analysis_STICKY_TRAPS_SPINCITY
Katrien (data from Henri, Silas and Nele)
2024-06-13
analysis prey communities - prey biomass - size distribution Sticky traps were sampled in 2021 Each location was succesfull trap was present for 1 week
#data setup
datainput cleanup exploration
here()## [1] "C:/Users/kadwolf/OneDrive - UGent/UGent-PC/Kadwolf/Documents/00_SPINCITY/henri/ST_mixed_model"set.seed(4)
data.gent<-read_csv2("data_ST_regio_Gent.csv", na=c("NB",""))
data.antwerpen<-read_csv2("data_ST_regio_Antwerpen.csv", na=c("NB",""))
data.leuven<-read_csv2("data_ST_regio_Leuven.csv", na=c("NB",""))
#check if colnames are the same can be done via names() or colnames(df)==colnames(data2)
#colnames(df_gent)==colnames(df_antwerpen)
#colnames(df_gent)==colnames(df_leuven)
#colnames(df_antwerpen)==colnames(df_leuven)
#combine the data and put all the variables correctly
data.st<-bind_rows(data.gent, data.antwerpen, data.leuven)
dim(data.st)## [1] 6386 20remove extra sampling P20SE, and rename columns
### remove unneccessary columns: limval type person ST_subdiv and rename to english
data.st <- data.st %>%
rename("sampleID"="staalid",
"order"= "orde",
"suborder"="onderorde",
"length"="lengte",
"width"="breedte",
"remark"="opmerking") %>%
select(start_date, end_date, sampleID, project, plotid, U_landscape, U_local, location,
phylum, class, order, suborder, length, width, n_indiv, remark)
data.st <- data.st %>%
filter(location != "P20SE")put all factors correctly
all 81 locations were sampled. P10SY and P13SY were less suitable (check if they are very different)
data.st$sampleID <- as.factor(data.st$sampleID)
data.st$project <- as.factor(data.st$project)
data.st$plotid <- as.factor(data.st$plotid)
levels(data.st$plotid) # 27 locations## [1] "P01" "P02" "P03" "P04" "P05" "P06" "P07" "P08" "P09" "P10" "P11" "P12"
## [13] "P13" "P14" "P15" "P16" "P17" "P18" "P19" "P20" "P21" "P22" "P23" "P24"
## [25] "P25" "P26" "P27"data.st$U_landscape <- as.factor(data.st$U_landscape)
data.st$U_landscape<- factor(data.st$U_landscape, levels=c("LOW", "MEDIUM", "HIGH"))
data.st$U_local <- as.factor(data.st$U_local)
data.st$U_local<- factor(data.st$U_local, levels=c("LOW", "MEDIUM", "HIGH"))
data.st$phylum <- factor(data.st$phylum)
data.st$class <- factor(data.st$class)
data.st$order <- factor(data.st$order)
data.st$suborder <- factor(data.st$suborder)
data.st$location <- factor(data.st$location)
data.st$location <- factor(data.st$location,c("P01SG", "P01SY", "P01SR", "P02SG", "P02SY", "P02SR", "P03SG", "P03SY", "P03SR", "P04SG", "P04SY", "P04SR", "P05SG", "P05SY", "P05SR", "P06SG", "P06SY", "P06SR", "P07SG", "P07SY", "P07SR", "P08SG", "P08SY", "P08SR", "P09SG", "P09SY", "P09SR", "P10SG", "P10SY", "P10SR", "P11SG", "P11SY", "P11SR", "P12SG", "P12SY", "P12SR", "P13SG", "P13SY", "P13SR", "P14SG", "P14SY", "P14SR", "P15SG", "P15SY", "P15SR", "P16SG", "P16SY", "P16SR", "P17SG", "P17SY", "P17SR", "P18SG", "P18SY", "P18SR", "P19SG", "P19SY", "P19SR", "P20SG", "P20SY", "P20SR", "P21SG", "P21SY", "P21SR", "P22SG", "P22SY", "P22SR", "P23SG", "P23SY", "P23SR", "P24SG", "P24SY", "P24SR", "P25SG", "P25SY", "P25SR", "P26SG", "P26SY", "P26SR", "P27SG", "P27SY", "P27SR"))
length(levels(data.st$location)) ## correct all 81 locations were successfully sampled## [1] 81data.st <- data.st %>%
mutate(urb_cat= paste0(U_landscape, U_local))
data.st$urb_cat <- as.factor(data.st$urb_cat)
data.st$urb_cat <- factor(data.st$urb_cat, levels=c("LOWLOW", "LOWMEDIUM", "LOWHIGH", "MEDIUMLOW", "MEDIUMMEDIUM", "MEDIUMHIGH", "HIGHLOW", "HIGHMEDIUM", "HIGHHIGH"))
urb.col9 <-c("#38A800", "lightgreen", "darkolivegreen3" , "yellow", "#E6E600", "goldenrod1", "deeppink3", "purple", "#8c008c") #colour vector for the 9 urbanisation categories
urb.col3<- c("#38A800", "#E6E600", "#8B008B") ##urbanisation colour vector to use in graphs
col_vec <- c(rep("#8B008B", 3), rep("#E6E600", 3), rep("#38A800", 3))
plotid.col27 <- rep(col_vec, 3) #handy to colour the facets boxes
data.st$day <- as.numeric(chron(dates. = as.character(data.st$start_date), format = c(dates = "d/m/y"), origin. = c(1, 1, 2021)))
## after this date can be put to a date object
data.st$start_date <- as.Date(data.st$start_date, format="%d/%m/%Y")
data.st$end_date <- as.Date(data.st$end_date, format="%d/%m/%Y")
glimpse(data.st)## Rows: 6,282
## Columns: 18
## $ start_date <date> 2021-09-07, 2021-09-07, 2021-09-07, 2021-09-07, 2021-09-0…
## $ end_date <date> 2021-09-14, 2021-09-14, 2021-09-14, 2021-09-14, 2021-09-1…
## $ sampleID <fct> SC21P01SRLASEP07, SC21P01SYLASEP07, SC21P01SYLASEP07, SC21…
## $ project <fct> SC21, SC21, SC21, SC21, SC21, SC21, SC21, SC21, SC21, SC21…
## $ plotid <fct> P01, P01, P01, P01, P01, P01, P01, P01, P01, P01, P01, P01…
## $ U_landscape <fct> HIGH, HIGH, HIGH, HIGH, HIGH, HIGH, HIGH, HIGH, HIGH, HIGH…
## $ U_local <fct> HIGH, MEDIUM, MEDIUM, LOW, LOW, HIGH, MEDIUM, LOW, LOW, LO…
## $ location <fct> P01SR, P01SY, P01SY, P01SG, P01SG, P01SR, P01SY, P01SG, P0…
## $ phylum <fct> ARTHROPODA, ARTHROPODA, ARTHROPODA, ARTHROPODA, ARTHROPODA…
## $ class <fct> ARACHNIDA, ARACHNIDA, ARACHNIDA, INSECTA, INSECTA, INSECTA…
## $ order <fct> ARANEAE, ARANEAE, ARANEAE, COLEOPTERA, COLEOPTERA, COLEOPT…
## $ suborder <fct> NA, NA, NA, POLYPHAGA, POLYPHAGA, POLYPHAGA, POLYPHAGA, BR…
## $ length <dbl> 1.5, 2.0, 0.8, 1.1, 0.8, 2.6, 1.8, 0.5, 7.4, 7.5, 7.5, 6.4…
## $ width <dbl> 1.3, 0.9, 0.4, 0.5, 0.5, 0.7, 0.9, 0.1, 3.5, 3.1, 2.7, 2.7…
## $ n_indiv <dbl> 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1…
## $ remark <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA…
## $ urb_cat <fct> HIGHHIGH, HIGHMEDIUM, HIGHMEDIUM, HIGHLOW, HIGHLOW, HIGHHI…
## $ day <dbl> 249, 249, 249, 249, 249, 249, 249, 249, 249, 249, 249, 249…names(which(colSums(is.na(data.st)) > 0)) ## this is okay## [1] "suborder" "remark"for some location both sticky traps were measured, resolve this by dividing n_indiv
## for some locations both sticky traps were measured and the average should be taken of the n_indiv
data.st<- data.st %>%
mutate(sampleID = case_when(
sampleID == "SC21P05SGLASEP07" ~ "SC21P05SGLBSEP07",
sampleID == "SC21P06SGLAAUG28" ~ "SC21P06SGLBAUG28",
sampleID == "SC21P07SYLASEP07" ~ "SC21P07SYLBSEP07",
sampleID == "SC21P08SGLAAUG18" ~ "SC21P08SGLBAUG18",
sampleID == "SC21P09SGLAAUG28" ~ "SC21P09SGLBAUG28",
sampleID == "SC21P10SGLASEP08" ~ "SC21P10SGLBSEP08",
sampleID == "SC21P11SGLAAUG27" ~ "SC21P11SGLBAUG27",
sampleID == "SC21P12SYLAAUG17" ~ "SC21P12SYLBAUG17",
sampleID == "SC21P15SYLAAUG27" ~ "SC21P15SYLBAUG27",
sampleID == "SC21P17SRLAAUG17" ~ "SC21P17SRLBAUG17",
TRUE ~ sampleID
))
double_st <- c("SC21P05SGLBSEP07", "SC21P06SGLBAUG28", "SC21P07SYLBSEP07", "SC21P08SGLBAUG18", "SC21P09SGLBAUG28", "SC21P10SGLBSEP08", "SC21P11SGLBAUG27", "SC21P12SYLBAUG17", "SC21P15SYLBAUG27", "SC21P17SRLBAUG17")
data.st <- data.st %>%
mutate(n_indiv=(ifelse(sampleID %in% double_st, ceiling(n_indiv/2), n_indiv)))next we will arrange the data in long so it is handy to use with ggplot
sum(data.st$n_indiv) #we want a dataset where there are as many rows as there are prey items measured, so equal to n_indiv## [1] 9605data.st <- data.st[rep(seq_len(nrow(data.st)), data.st$n_indiv), 1:ncol(data.st)]
nrow(data.st) # number of rows is now the same as the number of observations we only need to put the n_indiv to 1 ## [1] 9605data.st$n_indiv<-1
sum(data.st$n_indiv, na.rm=T) #double check if it is correct ok## [1] 9605calculate area of the insects the as measure for biomass
#create column with the surface area as a proxy of the biomass
#round the length measurements to 1 decimal
data.st$length <- round(data.st$length, 1)
data.st$width <- round(data.st$width, 1)
data.st <- data.st %>%
mutate(area=length*width)#data exploration order
avg1 <- data.st %>%
group_by(plotid, U_landscape, U_local, phylum, class, order) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T),
avg_length=mean(length,na.rm=T),
sd_avg_length = sd(length, na.rm=T))
avg2 <- data.st %>%
group_by(plotid, U_landscape, order) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T),
avg_length=mean(length,na.rm=T),
sd_avg_length = sd(length, na.rm=T))
# for each urb land and urb local combination
avg3 <- data.st %>%
group_by(U_landscape, U_local, order) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T),
avg_length=mean(length,na.rm=T),
sd_avg_length = sd(length, na.rm=T))
## U_local in columns
avg4 <-data.st %>%
group_by(U_landscape, U_local, order) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
pivot_wider(values_from = total_n_indiv, names_from=U_local)
## looks already like a kind of community matrix
avg5 <-data.st %>%
group_by(location, order) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
pivot_wider(values_from = total_n_indiv, names_from=location)
avg6 <-data.st %>%
group_by(U_landscape, U_local, order) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
pivot_wider(values_from = total_n_indiv, names_from=order)sub_order
avg1_sub <- data.st %>%
group_by(plotid, U_landscape, U_local, phylum, class, order, suborder) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T),
avg_length=mean(length,na.rm=T),
sd_avg_length = sd(length, na.rm=T))
avg2_sub <- data.st %>%
group_by(plotid, U_landscape, order, suborder) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T),
avg_length=mean(length,na.rm=T),
sd_avg_length = sd(length, na.rm=T))
# for each urb land and urb local combination
avg3_sub <- data.st %>%
group_by(U_landscape, U_local, order, suborder) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T),
avg_length=mean(length,na.rm=T),
sd_avg_length = sd(length, na.rm=T))
## U_local in columns
avg4_sub <-data.st %>%
group_by(U_landscape, U_local, order, suborder) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
pivot_wider(values_from = total_n_indiv, names_from=U_local)
## looks already like a kind of community matrix
avg5_sub <-data.st %>%
group_by(location, order, suborder) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
pivot_wider(values_from = total_n_indiv, names_from=location)prey abundance (number of specimens and relative abundance)
number of specimens for each location
data.st.tot <- data.st %>%
group_by(plotid, U_landscape, U_local, location) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T))
fig_n_indiv <- ggplot(data.st.tot, aes(x = U_local, y = total_n_indiv, fill=U_local)) +
geom_col() +
scale_fill_manual(values=urb.col3, guide="none")+
facet_grid(~plotid)+
ylab("number of specimens")+
xlab("urbanisation level at local scale")+
ggtitle("number of specimens across each sampling location")+
theme_bw()+
theme(axis.text.x = element_text(angle=45,hjust=1, size=rel(0.8)),
plot.title = element_text(size=rel(1.1)))
g <- ggplot_gtable(ggplot_build(fig_n_indiv))
stript <- which(grepl('strip-t', g$layout$name))
fills <- scales::alpha(plotid.col27, 0.30)
k <- 1
for (i in stript) {
j <- which(grepl('rect', g$grobs[[i]]$grobs[[1]]$childrenOrder))
g$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g)
data.st.tot[which.max(data.st.tot$total_n_indiv),]## # A tibble: 1 × 5
## # Groups: plotid, U_landscape, U_local [1]
## plotid U_landscape U_local location total_n_indiv
## <fct> <fct> <fct> <fct> <dbl>
## 1 P09 LOW LOW P09SG 487P09SG: largest number of specimens on the sticky trap
data.st.tot[which.min(data.st.tot$total_n_indiv),]## # A tibble: 1 × 5
## # Groups: plotid, U_landscape, U_local [1]
## plotid U_landscape U_local location total_n_indiv
## <fct> <fct> <fct> <fct> <dbl>
## 1 P24 MEDIUM LOW P24SG 15P24SG: smallest number of speciments on the sticky trap
##number of specimens along the urbanisation gradient
fig2 <- ggplot(data.st.tot, aes(x = U_local, y = total_n_indiv, color=U_local)) +
geom_point() +
scale_colour_manual(values=urb.col3, name="urbanisation level \nat local scale")+
facet_grid(~U_landscape)+
ylab("number of specimens")+
xlab("urbanisation at local scale")+
ggtitle("number of specimens along the urbanisation gradient")+
theme_bw()+
theme(axis.text = element_text(size=rel(1)), axis.title=element_text(size=rel(1)),
plot.title = element_text(size=rel(1.1)))
fig2
g <- ggplot_gtable(ggplot_build(fig2))
stript <- which(grepl('strip-t', g$layout$name))
fills <- scales::alpha(urb.col3, 0.30)
k <- 1
for (i in stript) {
j <- which(grepl('rect', g$grobs[[i]]$grobs[[1]]$childrenOrder))
g$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g)
same graph: other visualisation :
ggplot(data.st.tot, aes(U_landscape, total_n_indiv, fill= U_local)) +
geom_boxplot(position=position_dodge(width=0.75))+
geom_jitter(position=position_dodge(width=0.75), alpha=0.6, size=1.5, color="gray3")+
scale_fill_manual(values = urb.col3, name="urbanisation level \nat local scale") +
ylab("number of specimens")+
xlab("urbanisation level at landscape scale")+
ggtitle("number of specimens along the urbanisation gradient")+
theme_bw()+
theme(axis.text = element_text(size=rel(01)), axis.title=element_text(size=rel(1)))
explore order and suborder
##order ###by plotid graph showing the relative abundance of each arthropod order for each sampling location:
#16 ordes
col_ord <- c ("#323133", "#B8C0C7","#42A7F7", "blue4", "cyan", "goldenrod1", "#D9E33A", "yellow", "darkorchid4", "darkorchid1", "pink", "darkolivegreen1", "darkolivegreen4", "green", "firebrick1", "darkred" )
rel_n<- data.st %>%
group_by(plotid, U_landscape, U_local, location, order) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
ungroup()%>%
group_by(plotid, U_landscape, U_local, location) %>%
mutate(n_indiv_loc = sum(total_n_indiv, na.rm=T),
rel_n_indiv = total_n_indiv/n_indiv_loc ) %>%
ungroup()
### reorder the factors so it is plotted from highest number of specimens to lowest (more easy to compare)
rel_n$order <- reorder(rel_n$order, rel_n$rel_n_indiv)
rel_n$order <- factor(rel_n$order, levels=rev(levels(rel_n$order)))
## put also the order of data.st in this way (otherwise the colour code will not be preserved for each order
data.st$order <- factor(data.st$order, levels=levels(rel_n$order))
fig_rel_ord<- rel_n %>% ggplot(aes(x = U_local, y = rel_n_indiv, fill=order)) +
geom_col() +
scale_fill_manual(values=col_ord)+
facet_grid(~plotid)+
ylab("relative abundance")+
scale_y_continuous(labels=scales::percent_format())+
xlab("urbanisation level at local scale")+
ggtitle("relative abundance of each taxonomic order for each sampling location")+
theme_bw()+
theme(axis.text.x = element_text(angle=45,hjust=1, size=rel(0.8)),
plot.title = element_text(size=rel(1.1)))
g <- ggplot_gtable(ggplot_build(fig_rel_ord))
stript <- which(grepl('strip-t', g$layout$name))
fills <- scales::alpha(plotid.col27, 0.30)
k <- 1
for (i in stript) {
j <- which(grepl('rect', g$grobs[[i]]$grobs[[1]]$childrenOrder))
g$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g)
graph showing the number of specimens of each arthropod order for each sampling location:
ord_n <- data.st %>%
group_by(plotid, U_landscape, U_local, location, order) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
ungroup()
fig_ord<- ord_n %>% ggplot(aes(x = U_local, y = total_n_indiv, fill=order)) +
geom_col() +
scale_fill_manual(values=col_ord)+
facet_grid(~plotid)+
ylab("number of specimens")+
xlab("urbanisation level at local scale")+
ggtitle("number of specimens of each taxonomic order for each sampling location")+
theme_bw()+
theme(axis.text.x = element_text(angle=45,hjust=1, size=rel(0.8)),
plot.title = element_text(size=rel(1.1)))
g <- ggplot_gtable(ggplot_build(fig_ord))
stript <- which(grepl('strip-t', g$layout$name))
fills <- scales::alpha(plotid.col27, 0.30)
k <- 1
for (i in stript) {
j <- which(grepl('rect', g$grobs[[i]]$grobs[[1]]$childrenOrder))
g$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g)
by urb level
the relative abundance of the Arthropod orders across different urbanisation levels at landscape and local scale:
rel_n_urb<- data.st %>%
group_by(U_landscape, U_local, urb_cat, order) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
ungroup()%>%
group_by(U_landscape, U_local, urb_cat) %>%
mutate(n_indiv_urb = sum(total_n_indiv, na.rm=T),
rel_n_indiv = total_n_indiv/n_indiv_urb ) %>%
ungroup()
fig_rel_ord_urb<- rel_n_urb %>% ggplot(aes(x = U_local, y = rel_n_indiv, fill=order)) +
geom_col() +
scale_fill_manual(values=col_ord)+
facet_grid(~U_landscape)+
ylab("relative abundance")+
scale_y_continuous(labels=scales::percent_format())+
xlab("urbanisation level at local scale")+
ggtitle("relative abundance of the taxonomic orders sampled along the urbanisation gradient")+
theme_bw()+
theme(axis.text.x = element_text(size=rel(1)),
plot.title = element_text(size=rel(1.1)))
g <- ggplot_gtable(ggplot_build(fig_rel_ord_urb))
stript <- which(grepl('strip-t', g$layout$name))
fills <- scales::alpha(urb.col3, 0.30)
k <- 1
for (i in stript) {
j <- which(grepl('rect', g$grobs[[i]]$grobs[[1]]$childrenOrder))
g$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g)
graph showing the number of specimens of each arthropod order across different urbanisation levels at landscape and local scale:
n_urb<- data.st %>%
group_by(U_landscape, U_local, urb_cat, order) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
ungroup()
fig_n_ord_urb<- n_urb %>% ggplot(aes(x = U_local, y = total_n_indiv, fill=order)) +
geom_col() +
scale_fill_manual(values=col_ord)+
facet_grid(~U_landscape)+
ylab("number of specimens")+
xlab("urbanisation level at local scale")+
ggtitle("number of specimens of the taxonomic orders sampled along the urbanisation gradient")+
theme_bw()+
theme(axis.text.x = element_text(size=rel(1)),
plot.title = element_text(size=rel(1.1)))
g <- ggplot_gtable(ggplot_build(fig_n_ord_urb))
stript <- which(grepl('strip-t', g$layout$name))
fills <- scales::alpha(urb.col3, 0.30)
k <- 1
for (i in stript) {
j <- which(grepl('rect', g$grobs[[i]]$grobs[[1]]$childrenOrder))
g$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g)
##top3 orders diptera hemiptera hymenoptera are the top 3 orders
top3 <- c("DIPTERA", "HEMIPTERA", "HYMENOPTERA")###by plotid graph showing the relative abundance of the top3 arthropod orders for each sampling location:
top3_rel_n<- data.st %>% filter(order %in% top3) %>%
group_by(plotid, U_landscape, U_local, location, order) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
ungroup()%>%
group_by(plotid, U_landscape, U_local, location) %>%
mutate(n_indiv_loc = sum(total_n_indiv, na.rm=T),
rel_n_indiv = total_n_indiv/n_indiv_loc ) %>%
ungroup()
fig_top3_rel_ord<- top3_rel_n %>% ggplot(aes(x = U_local, y = rel_n_indiv, fill=order)) +
geom_col() +
scale_fill_manual(values=col_ord)+
facet_grid(~plotid)+
ylab("relative abundance")+
scale_y_continuous(labels=scales::percent_format())+
xlab("urbanisation level at local scale")+
ggtitle("relative abundance of the top3 athropod orders for each sampling location")+
theme_bw()+
theme(axis.text.x = element_text(angle=45,hjust=1, size=rel(0.8)),
plot.title = element_text(size=rel(1.1)))
g <- ggplot_gtable(ggplot_build(fig_top3_rel_ord))
stript <- which(grepl('strip-t', g$layout$name))
fills <- scales::alpha(plotid.col27, 0.30)
k <- 1
for (i in stript) {
j <- which(grepl('rect', g$grobs[[i]]$grobs[[1]]$childrenOrder))
g$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g)
graph showing the number of specimens : graph showing the number of specimens of the top3 arthropod orders for each sampling location:
top3_ord_n <- data.st %>% filter(order %in% top3) %>%
group_by(plotid, U_landscape, U_local, location, order) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
ungroup()
fig_top3_ord<- top3_ord_n %>% ggplot(aes(x = U_local, y = total_n_indiv, fill=order)) +
geom_col() +
scale_fill_manual(values=col_ord)+
facet_grid(~plotid)+
ylab("number of specimens")+
xlab("urbanisation level at local scale")+
ggtitle("number of specimens of the top3 arthropod orders for each location")+
theme_bw()+
theme(axis.text.x = element_text(angle=45,hjust=1, size=rel(0.8)),
plot.title = element_text(size=rel(1.1)))
g <- ggplot_gtable(ggplot_build(fig_top3_ord))
stript <- which(grepl('strip-t', g$layout$name))
fills <- scales::alpha(plotid.col27, 0.30)
k <- 1
for (i in stript) {
j <- which(grepl('rect', g$grobs[[i]]$grobs[[1]]$childrenOrder))
g$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g)
###by urb level graph showing the relative abundance : the relative abundance of the top3 Arthropod orders across different urbanisation levels at landscape and local scale
top3_rel_n_urb<- data.st %>% filter(order %in% top3) %>%
group_by(U_landscape, U_local, urb_cat, order) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
ungroup()%>%
group_by(U_landscape, U_local, urb_cat) %>%
mutate(n_indiv_urb = sum(total_n_indiv, na.rm=T),
rel_n_indiv = total_n_indiv/n_indiv_urb ) %>%
ungroup()
fig_top3_rel_ord_urb<- top3_rel_n_urb %>% ggplot(aes(x=U_local, y=rel_n_indiv, fill=order)) +
geom_col() +
scale_fill_manual(values=col_ord)+
facet_grid(~U_landscape)+
ylab("relative abundance")+
scale_y_continuous(labels=scales::percent_format())+
xlab("urbanisation level at local scale")+
ggtitle("relative abundance of the top3 Arthropod orders sampled along the urbanisation gradient")+
theme_bw()+
theme(axis.text.x = element_text(size=rel(1)),
plot.title = element_text(size=rel(1.1)))
g <- ggplot_gtable(ggplot_build(fig_top3_rel_ord_urb))
stript <- which(grepl('strip-t', g$layout$name))
fills <- scales::alpha(urb.col3, 0.30)
k <- 1
for (i in stript) {
j <- which(grepl('rect', g$grobs[[i]]$grobs[[1]]$childrenOrder))
g$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g)
the same graph but with the number of specimens : the number of specimens of the top3 Arthropod orders across different urbanisation levels at landscape and local scale
top3_n_urb<- data.st %>% filter(order %in% top3) %>%
group_by(U_landscape, U_local, urb_cat, order) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
ungroup()
fig_top3_n_ord_urb<- top3_n_urb %>% ggplot(aes(x = U_local, y = total_n_indiv, fill=order)) +
geom_col() +
scale_fill_manual(values=col_ord)+
facet_grid(~U_landscape)+
ylab("number of specimens")+
xlab("urbanisation level at local scale")+
ggtitle("number of specimens of the top3 Arthropod orders sampled along the urbanisation gradient")+
theme_bw()+
theme(axis.text.x = element_text(size=rel(1)),
plot.title = element_text(size=rel(1.1)))
g <- ggplot_gtable(ggplot_build(fig_top3_n_ord_urb))
stript <- which(grepl('strip-t', g$layout$name))
fills <- scales::alpha(urb.col3, 0.30)
k <- 1
for (i in stript) {
j <- which(grepl('rect', g$grobs[[i]]$grobs[[1]]$childrenOrder))
g$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g)
suborder - top3 orders
investigate suborder within the top3 orders
by plotid
the number of specimens of the suborders within the top3 arthropod orders across different urbanisation levels at landscape and local scale
### for certain orders the suborder is not known so NA
## however for diptera, hemiptera and hymenoptera the data is complete
s1 <- data.st %>%
filter(order=="DIPTERA") %>%
group_by(plotid, U_landscape, U_local, suborder) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T))%>%
ungroup()%>%
mutate(order="DIPTERA")
s2 <- data.st %>%
filter(order=="HEMIPTERA") %>%
group_by(plotid, U_landscape, U_local, suborder) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
ungroup()%>%
mutate(order="HEMIPTERA")
s3 <- data.st %>%
filter(order=="HYMENOPTERA") %>%
group_by(plotid, U_landscape, U_local, suborder) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
ungroup()%>%
mutate(order="HYMENOPTERA") %>%
ungroup()
s <- rbind(s1, s2, s3) ## number of specimens of the suborders within the top3 orders
fig_s<- s %>%
ggplot(aes(x = U_local, y = total_n_indiv, fill=suborder)) +
geom_bar(stat="identity") +
scale_fill_viridis_d()+
facet_grid(order~plotid)+
ylab("number of specimens")+
xlab("urbanisation level at local scale")+
ggtitle("number of specimens per suborder of the top 3 orders for each location")+
theme_bw()+
theme(legend.position="bottom",
axis.text.x = element_text(angle=45,hjust=1, size=rel(0.7)),
plot.title = element_text(size=rel(1.1)))
col_ord_3 <- col_ord[1:3]
g1 <- ggplot_gtable(ggplot_build(fig_s))
strip <- which(grepl('strip-', g1$layout$name))
fills <- c(scales::alpha(plotid.col27, 0.30), scales::alpha(col_ord_3, 0.30))
k <- 1
for (i in strip) {
j <- which(grepl('rect', g1$grobs[[i]]$grobs[[1]]$childrenOrder))
g1$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g1)
the relative abundance of suborders within the top3 arthropod orders across different urbanisation levels at landscape and local scale:
r1<- data.st %>%
filter(order=="DIPTERA") %>%
group_by(plotid, U_landscape, U_local, suborder) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T))%>%
ungroup()%>%
group_by(plotid, U_landscape, U_local) %>%
mutate(n_indiv_loc = sum(total_n_indiv, na.rm=T),
rel_n_indiv = total_n_indiv/n_indiv_loc,
order="DIPTERA") %>%
ungroup()
r2<- data.st %>%
filter(order=="HEMIPTERA") %>%
group_by(plotid, U_landscape, U_local, suborder) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T))%>%
ungroup()%>%
group_by(plotid, U_landscape, U_local) %>%
mutate(n_indiv_loc = sum(total_n_indiv, na.rm=T),
rel_n_indiv = total_n_indiv/n_indiv_loc,
order="HEMIPTERA") %>%
ungroup()
r3<- data.st %>%
filter(order=="HYMENOPTERA") %>%
group_by(plotid, U_landscape, U_local, suborder) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T))%>%
ungroup()%>%
group_by(plotid, U_landscape, U_local) %>%
mutate(n_indiv_loc = sum(total_n_indiv, na.rm=T),
rel_n_indiv = total_n_indiv/n_indiv_loc,
order="HYMENOPTERA") %>%
ungroup()
r <- rbind(r1, r2, r3) ## relative abundance of the suborders within the top3 orders
fig_r <- r %>%
ggplot(aes(x = U_local, y = rel_n_indiv, fill=suborder)) +
geom_bar(stat="identity") +
scale_fill_viridis_d()+
facet_grid(order~plotid)+
ylab("relative abundance")+
scale_y_continuous(labels=scales::percent_format())+
xlab("urbanisation level at local scale")+
ggtitle("relative abundance of the suborders within the top 3 taxonomic orders for each sampling location")+
theme_bw()+
theme(legend.position="bottom",
axis.text.x = element_text(angle=45,hjust=1, size=rel(0.7)),
plot.title = element_text(size=rel(1.1)))
col_ord_3 <- col_ord[1:3]
g1 <- ggplot_gtable(ggplot_build(fig_r))
strip <- which(grepl('strip-', g1$layout$name))
fills <- c(scales::alpha(plotid.col27, 0.30), scales::alpha(col_ord_3, 0.30))
k <- 1
for (i in strip) {
j <- which(grepl('rect', g1$grobs[[i]]$grobs[[1]]$childrenOrder))
g1$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g1)
by urb level
relative abundance of the suborders within the top3 orders for the different urbanisation levels at landscape and local scale:
### for certain orders the suborder is not known so NA
## however for diptera, hemiptera and hymenoptera the data is complete
rel_sub_1 <- data.st %>%
filter(order=="DIPTERA") %>%
group_by(U_landscape, U_local, urb_cat, suborder) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
ungroup()%>%
group_by(U_landscape, U_local, urb_cat) %>%
mutate(n_indiv_loc = sum(total_n_indiv, na.rm=T),
rel_n_indiv = total_n_indiv/n_indiv_loc,
order="DIPTERA") %>%
ungroup()
rel_sub_2 <- data.st %>%
filter(order=="HEMIPTERA") %>%
group_by(U_landscape, U_local, urb_cat, suborder) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
ungroup()%>%
group_by(U_landscape, U_local, urb_cat) %>%
mutate(n_indiv_loc = sum(total_n_indiv, na.rm=T),
rel_n_indiv = total_n_indiv/n_indiv_loc,
order="HEMIPTERA") %>%
ungroup()
rel_sub_3 <- data.st %>%
filter(order=="HYMENOPTERA") %>%
group_by(U_landscape, U_local, urb_cat, suborder) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
ungroup()%>%
group_by(U_landscape, U_local, urb_cat) %>%
mutate(n_indiv_loc = sum(total_n_indiv, na.rm=T),
rel_n_indiv = total_n_indiv/n_indiv_loc,
order="HYMENOPTERA") %>%
ungroup()
rel_sub <- rbind(rel_sub_1, rel_sub_2, rel_sub_3)
fig_rel_subord<-rel_sub %>%
ggplot(aes(x = U_local, y = rel_n_indiv, fill=suborder)) +
geom_bar(stat="identity") +
scale_fill_viridis_d()+
facet_grid(order~U_landscape)+
ylab("relative abundance")+
scale_y_continuous(labels=scales::percent_format())+
xlab("urbanisation level at local scale")+
ggtitle("relative abundance of the suborders within the top 3 orders \nacross different urbanisation levels at landscape and local scale")+
theme_bw()+
theme(plot.title = element_text(size=rel(1.1)))
col_ord_3 <- col_ord[1:3]
g1 <- ggplot_gtable(ggplot_build(fig_rel_subord))
strip <- which(grepl('strip-', g1$layout$name))
fills <- c(scales::alpha(urb.col3, 0.30), scales::alpha(col_ord_3, 0.30))
k <- 1
for (i in strip) {
j <- which(grepl('rect', g1$grobs[[i]]$grobs[[1]]$childrenOrder))
g1$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g1)
number of specimens of the suborders within the top3 orders for the different urbanisation levels at landscape and local scale
### for certain orders the suborder is not known so NA
## however for diptera, hemiptera and hymenoptera the data is complete
sub_1 <- data.st %>%
filter(order=="DIPTERA") %>%
group_by(U_landscape, U_local, urb_cat, suborder) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T))%>%
ungroup()%>%
mutate(order="DIPTERA")
sub_2 <- data.st %>%
filter(order=="HEMIPTERA") %>%
group_by(U_landscape, U_local, urb_cat, suborder) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
ungroup()%>%
mutate(order="HEMIPTERA")
sub_3 <- data.st %>%
filter(order=="HYMENOPTERA") %>%
group_by(U_landscape, U_local, urb_cat, suborder) %>%
summarise(total_n_indiv=sum(n_indiv, na.rm=T)) %>%
ungroup()%>%
mutate(order="HYMENOPTERA") %>%
ungroup()
sub <- rbind(sub_1, sub_2, sub_3)
fig_subord<- sub %>%
ggplot(aes(x = U_local, y = total_n_indiv, fill=suborder)) +
geom_bar(stat="identity") +
scale_fill_viridis_d()+
facet_grid(order~U_landscape)+
ylab("number of specimens")+
xlab("urbanisation level at local scale")+
ggtitle("number of specimens per suborder of the top 3 orders for each urbanisation level")+
theme_bw()+
theme(plot.title = element_text(size=rel(1.1)))
col_ord_3 <- col_ord[1:3]
g1 <- ggplot_gtable(ggplot_build(fig_subord))
strip <- which(grepl('strip-', g1$layout$name))
fills <- c(scales::alpha(urb.col3, 0.30), scales::alpha(col_ord_3, 0.30))
k <- 1
for (i in strip) {
j <- which(grepl('rect', g1$grobs[[i]]$grobs[[1]]$childrenOrder))
g1$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g1)
summary_top3 <- data.st%>%
filter(order %in% top3) %>%
droplevels()%>%
group_by(plotid, U_landscape, U_local, order, suborder) %>%
mutate(total_n_indiv=sum(n_indiv, na.rm=T))
### summary of the data with the total number of individuals per order)#prey size distribution
##length
p<- ggplot(data.st, aes(x = U_local, y = length, color=U_local)) +
geom_point() +
scale_color_manual(values=urb.col3, name="urbanisation level at local scale")+
facet_grid(~plotid)+
ylab("length (mm)")+
theme(legend.position="bottom", axis.title.x=element_blank(),
axis.text.x = element_blank(), axis.ticks.x=element_blank(),
plot.title = element_text(size=rel(1.1)))
g <- ggplot_gtable(ggplot_build(p))
stript <- which(grepl('strip-t', g$layout$name))
fills <- scales::alpha(plotid.col27, 0.30)
k <- 1
for (i in stript) {
j <- which(grepl('rect', g$grobs[[i]]$grobs[[1]]$childrenOrder))
g$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g)
most individuals are small
## same histogram with ggplot
lenghist1 <- ggplot(data.st, aes(x=length)) +
geom_histogram(binwidth=0.5) +
ggtitle("length per 0.5 mm") +
theme_bw()
lenghist2 <- ggplot(data.st, aes(x=length)) +
geom_histogram(binwidth=0.25) +
ggtitle("length per 0.25 mm") +
theme_bw()
lenghist3 <- ggplot(data.st, aes(x=length)) +
geom_histogram(binwidth=0.1)+
ggtitle("length per 0.10 mm") +
theme_bw()
ggarrange(lenghist1, lenghist2, lenghist3, ncol = 3)
area
areahist1 <- ggplot(data.st, aes(x=area)) +
geom_histogram(binwidth=1)+
ggtitle(bquote('area hist 1 ('~mm^2~')')) +
theme_bw()
areahist2 <- ggplot(data.st, aes(x=area)) +
geom_histogram(binwidth=0.1)+
ggtitle(bquote('area hist 0.1 ('~mm^2~')')) +
theme_bw()
areahist3 <- ggplot(data.st, aes(x=area)) +
geom_histogram(binwidth=0.05)+
ggtitle(bquote('area hist 0.05 ('~mm^2~')')) +
theme_bw()
ggarrange(areahist1, areahist2, areahist3, ncol = 3)
maybe best to use the length to divide the specimens in size classes? if we take all samples it looks like 0.1mm presents the best => but if we divide the samples in their urbanisation categories 0.5 or even 1 could work => then we have 45 size classes (otherwise 225)
hist1 <- ggplot(data.st, aes(x=length)) +
geom_histogram(binwidth=0.1)+
ggtitle("histogram: length per 0.1 mm") +
facet_grid(U_landscape~U_local)+
scale_fill_manual(values=urb.col3)+
theme_bw()
hist1
g1 <- ggplot_gtable(ggplot_build(hist1))
strip <- which(grepl('strip-', g1$layout$name))
fills <- c(scales::alpha(urb.col3, 0.30), scales::alpha(urb.col3, 0.30))
k <- 1
for (i in strip) {
j <- which(grepl('rect', g1$grobs[[i]]$grobs[[1]]$childrenOrder))
g1$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g1)
hist2 <- ggplot(data.st, aes(x=length)) +
geom_histogram(binwidth=0.5)+
ggtitle("histogram; length per 0.5 mm") +
facet_grid(U_landscape~U_local)+
scale_fill_manual(values=urb.col3)+
theme_bw()
hist2
g2 <- ggplot_gtable(ggplot_build(hist2))
strip <- which(grepl('strip-', g1$layout$name))
fills <- c(scales::alpha(urb.col3, 0.30), scales::alpha(urb.col3, 0.30))
k <- 1
for (i in strip) {
j <- which(grepl('rect', g2$grobs[[i]]$grobs[[1]]$childrenOrder))
g2$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g2)
hist3 <- ggplot(data.st, aes(x=length)) +
geom_histogram(binwidth=1)+
ggtitle("histogram: length per 1 mm") +
facet_grid(U_landscape~U_local)+
scale_fill_manual(values=urb.col3)+
theme_bw()
hist3
g3 <- ggplot_gtable(ggplot_build(hist3))
strip <- which(grepl('strip-', g1$layout$name))
fills <- c(scales::alpha(urb.col3, 0.30), scales::alpha(urb.col3, 0.30))
k <- 1
for (i in strip) {
j <- which(grepl('rect', g3$grobs[[i]]$grobs[[1]]$childrenOrder))
g3$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g3)
ggarrange(g1, g2, g3, nrow = 3)
##functional groups/species maybe work with 0.5 mm (or if necessary with 1mm but maybe this is to wide of a range). if we continue with 0.5 we have 45 size classes
breaks <-c(seq(0,22, by=0.5), Inf)
labels <- paste0("SC", 1:(length(breaks)-1))
labels <- c("SC01","SC02","SC03","SC04","SC05","SC06","SC07","SC08","SC09","SC10","SC11","SC12","SC13","SC14","SC15","SC16","SC17","SC18","SC19","SC20","SC21","SC22","SC23", "SC24", "SC25", "SC26", "SC27", "SC28", "SC29", "SC30", "SC31", "SC32", "SC33", "SC34", "SC35", "SC36", "SC37", "SC38", "SC39", "SC40", "SC41", "SC42", "SC43", "SC44", "SC45")
data.st$sizeclass <- cut(data.st$length, breaks = breaks, labels = labels, right = FALSE)
#then merge the data of the size classes with the taxonomy data
data.st$suborder <- ifelse(is.na(data.st$suborder), "ord", paste(data.st$suborder))
data.st <- data.st %>%
mutate(ord_sub = paste(str_sub(order, 1, 3), str_sub(suborder, 1, 3), sep="_"))
data.st <- data.st %>%
mutate(func_species = paste(sizeclass, ord_sub, sep="_"))
data.st$ord_sub <- as.factor(data.st$ord_sub)
length(levels(data.st$ord_sub)) ## [1] 22data.st$func_species <- as.factor(data.st$func_species)
length(levels(data.st$func_species)) # in total 176 functional species## [1] 197in total 197 functional species
FUNC GROUPS IN top3 orders THE DIPTERA-HYMENOPTERA-HEMIPTERA
sub_dip_col <- c("#31688EFF","#35B779FF")
sub_hem_col <- c("#443A83FF", "#21908CFF", "#8FD744FF")
sub_hym_col <- c("#440154FF", "#FDE725FF")d<-data.st %>%
filter(order == "DIPTERA") %>%
droplevels() %>%
ggplot(aes(x = U_local, y = n_indiv, fill=func_species)) +
geom_bar(stat="identity") +
scale_fill_viridis_d()+
theme(axis.text.x = element_text(angle=45,hjust=1)) +
ggtitle("DIPTERA functional groups") +
facet_grid(suborder~U_landscape)+
ylab("number of specimens per functional species")
g1 <- ggplot_gtable(ggplot_build(d))
strip <- which(grepl('strip-', g1$layout$name))
fills <- c(scales::alpha(urb.col3, 0.30), scales::alpha(sub_dip_col, 0.50))
k <- 1
for (i in strip) {
j <- which(grepl('rect', g1$grobs[[i]]$grobs[[1]]$childrenOrder))
g1$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g1)
d<-data.st %>%
filter(order == "HEMIPTERA") %>%
droplevels() %>%
ggplot(aes(x = U_local, y = n_indiv, fill=func_species)) +
geom_bar(stat="identity") +
scale_fill_viridis_d()+
theme(axis.text.x = element_text(angle=45,hjust=1)) +
ggtitle("HEMIPTERA functional groups") +
facet_grid(suborder~U_landscape)+
ylab("number of specimens per functional species")
g1 <- ggplot_gtable(ggplot_build(d))
strip <- which(grepl('strip-', g1$layout$name))
fills <- c(scales::alpha(urb.col3, 0.30), scales::alpha(sub_hem_col, 0.50))
k <- 1
for (i in strip) {
j <- which(grepl('rect', g1$grobs[[i]]$grobs[[1]]$childrenOrder))
g1$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g1)
d<-data.st %>%
filter(order == "HYMENOPTERA") %>%
droplevels() %>%
ggplot(aes(x = U_local, y = n_indiv, fill=func_species)) +
geom_bar(stat="identity") +
scale_fill_viridis_d()+
theme(axis.text.x = element_text(angle=45,hjust=1)) +
ggtitle("HYMENOPTERA functional groups") +
facet_grid(suborder~U_landscape)+
ylab("number of specimens per functional species")
g1 <- ggplot_gtable(ggplot_build(d))
strip <- which(grepl('strip-', g1$layout$name))
fills <- c(scales::alpha(urb.col3, 0.30), scales::alpha(sub_hym_col, 0.50))
k <- 1
for (i in strip) {
j <- which(grepl('rect', g1$grobs[[i]]$grobs[[1]]$childrenOrder))
g1$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g1)
dark blue or low size class number => small size individuals
statistics
#standardize the (continuous) explanatory variables so they have a mean of zero and standard deviation of one, so the estimated coefficients are all on the same scale and you can compare more easy the effect sizes
data.st$sday <- scale(data.st$day)
#data.st$slengte <- scale(data.st$length)
#data.st$sopp <- scale(data.st$area)
#assumptions mixed model: look at the residuals (they need to be +- normally distributed) plot(mixed.model, which=1 or 2) plot the residuals - the red line should be close to being flat, like the dashed grey linenumber of specimens
poisson => but dharma test gives some strange results
no indication of correlation with sampling day U_local and U_landscape significant
df_N <- data.st %>%
group_by(plotid, U_landscape, U_local, urb_cat, sday, day) %>%
mutate(total_n_indiv=sum(n_indiv, na.rm=T)) %>% ungroup()
df <- df_N %>%
select(total_n_indiv, day) %>% ungroup()
df_cor<-cor(df, method = "pearson")
corrplot(df_cor,method = "circle", addgrid.col="grey", addCoef.col = "black", number.cex = 0.8, type="upper", title="correlation number of specimens and sampling day", mar=c(0,0,2,0))
df_N$U_landscape <- as.factor(df_N$U_landscape)
df_N$U_landscape<- factor(df_N$U_landscape, levels=c("LOW", "MEDIUM", "HIGH"))
df_N$U_local <- as.factor(df_N$U_local)
df_N$U_local<- factor(df_N$U_local, levels=c("LOW", "MEDIUM", "HIGH"))
modA<-glmmTMB(total_n_indiv ~ U_landscape*U_local+(1|plotid), data=df_N, family=poisson) #nbinom 2 to correct for overdispersion
summary(modA)## Family: poisson ( log )
## Formula: total_n_indiv ~ U_landscape * U_local + (1 | plotid)
## Data: df_N
##
## AIC BIC logLik deviance df.resid
## 249584.8 249656.5 -124782.4 249564.8 9595
##
## Random effects:
##
## Conditional model:
## Groups Name Variance Std.Dev.
## plotid (Intercept) 0.1897 0.4356
## Number of obs: 9605, groups: plotid, 27
##
## Conditional model:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 5.139872 0.145211 35.40 <2e-16 ***
## U_landscapeMEDIUM -0.304649 0.205369 -1.48 0.138
## U_landscapeHIGH -0.140724 0.205364 -0.69 0.493
## U_localMEDIUM -0.228634 0.002851 -80.21 <2e-16 ***
## U_localHIGH -0.431410 0.003406 -126.68 <2e-16 ***
## U_landscapeMEDIUM:U_localMEDIUM 0.105811 0.004781 22.13 <2e-16 ***
## U_landscapeHIGH:U_localMEDIUM -0.107727 0.005177 -20.81 <2e-16 ***
## U_landscapeMEDIUM:U_localHIGH 0.301714 0.004892 61.67 <2e-16 ***
## U_landscapeHIGH:U_localHIGH 0.291675 0.005096 57.24 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Anova(modA, type="3") #U_local and U_landscape is significant## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: total_n_indiv
## Chisq Df Pr(>Chisq)
## (Intercept) 1252.8705 1 <2e-16 ***
## U_landscape 2.2048 2 0.3321
## U_local 17325.0310 2 <2e-16 ***
## U_landscape:U_local 7005.2389 4 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot(simulateResiduals(modA))
testDispersion(simulateResiduals(modA))
##
## DHARMa nonparametric dispersion test via sd of residuals fitted vs.
## simulated
##
## data: simulationOutput
## dispersion = 2.0615, p-value = 0.064
## alternative hypothesis: two.sidedplot(allEffects(modA))
contrast(emmeans(modA, specs=~U_local), method="pairwise") |> as_tibble() ## # A tibble: 3 × 6
## contrast estimate SE df z.ratio p.value
## <chr> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 LOW - MEDIUM 0.229 0.00215 Inf 107. 0
## 2 LOW - HIGH 0.234 0.00206 Inf 113. 0
## 3 MEDIUM - HIGH 0.00434 0.00237 Inf 1.83 0.160modA_comp <- contrast(emmeans(modA, specs=~U_landscape*U_local), method="pairwise") |> as_tibble()
modA_comp %>% filter(p.value < 0.099) %>% knitr::kable() | contrast | estimate | SE | df | z.ratio | p.value |
|---|---|---|---|---|---|
| LOW LOW - LOW MEDIUM | 0.2286342 | 0.0028505 | Inf | 80.20771 | 0 |
| LOW LOW - LOW HIGH | 0.4314100 | 0.0034055 | Inf | 126.67941 | 0 |
| MEDIUM LOW - MEDIUM MEDIUM | 0.1228229 | 0.0038389 | Inf | 31.99454 | 0 |
| MEDIUM LOW - MEDIUM HIGH | 0.1296961 | 0.0035122 | Inf | 36.92762 | 0 |
| HIGH LOW - HIGH MEDIUM | 0.3363613 | 0.0043218 | Inf | 77.82828 | 0 |
| HIGH LOW - HIGH HIGH | 0.1397353 | 0.0037906 | Inf | 36.86351 | 0 |
| LOW MEDIUM - LOW HIGH | 0.2027758 | 0.0035379 | Inf | 57.31556 | 0 |
| HIGH MEDIUM - HIGH HIGH | -0.1966259 | 0.0045611 | Inf | -43.10957 | 0 |
preds_modA <- emmeans(modA, specs=~U_landscape*U_local, type="response") |> as_tibble()
preds_modA## # A tibble: 9 × 7
## U_landscape U_local rate SE df asymp.LCL asymp.UCL
## <fct> <fct> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 LOW LOW 171. 24.8 Inf 128. 227.
## 2 MEDIUM LOW 126. 18.3 Inf 94.7 167.
## 3 HIGH LOW 148. 21.5 Inf 112. 197.
## 4 LOW MEDIUM 136. 19.7 Inf 102. 181.
## 5 MEDIUM MEDIUM 111. 16.2 Inf 83.7 148.
## 6 HIGH MEDIUM 106. 15.4 Inf 79.7 141.
## 7 LOW HIGH 111. 16.1 Inf 83.4 147.
## 8 MEDIUM HIGH 111. 16.1 Inf 83.2 147.
## 9 HIGH HIGH 129. 18.7 Inf 97.0 171.pd <- position_dodge(width=0.5)
g_N<-ggplot()+
geom_point(data=df_N,
aes(x=U_landscape, y=total_n_indiv,color=U_local),
position=pd, size=2)+
geom_pointrange(data=preds_modA,
aes(x=U_landscape, y=rate, ymin=asymp.LCL, ymax=asymp.UCL,
fill=U_local),
position=pd, size = 1, pch = 21)+
scale_fill_manual(values=urb.col3, name="Urbanisation level \nat local scale")+
scale_colour_manual(values=urb.col3, guide="none")+
labs(x="urbanisation level at landscape scale", y="number of specimens")+
ggtitle("effect of urbanisation on arthropod abundance")+
theme_bw()+
theme(axis.text = element_text(size=rel(1)),
axis.title = element_text(size=rel(1.1)),
plot.title = element_text(size=rel(1.1)))
g_N
biomass
random factor : (1|plotid) fixed effects : urbanisation at landscape and local scale :U_landscape and U_local
sday : scaled sampling day test but did not influence the biomass data
calculated biomass as : W = a * L^b describing how body length (L) and biomass is correlated
a = 0.04 b = 2.26 source: Sabo et al. 2002: Sabo, J.L., Bastow, J.L. & Power, M.E., 2002. Length-Mass Relationships for Adult Aquatic and Terrestrial Invertebrates in a California Watershed.
(in study of the netherlands they us a very general a=0.03 b=2.63)
effect of urbanisation on arthropod biomass
data.st <- data.st %>%
mutate(biomass_indiv = 0.04*(length^2.26))
df_biomass <- data.st %>%
group_by(sampleID, plotid, U_landscape, U_local, urb_cat, day, sday) %>%
summarise(biomass = sum(biomass_indiv))%>% ungroup()
df <- df_biomass %>%
select(biomass, day) %>% ungroup()
df_cor<-cor(df, method = "pearson")
corrplot(df_cor,method = "circle", addgrid.col="grey", addCoef.col = "black", number.cex = 0.8, type="upper", title="correlation biomass and sampling day", mar=c(0,0,2,0))
modA<-glmmTMB(biomass ~ U_landscape+U_local+ (1|plotid),data=df_biomass, family=Gamma(link=log))
summary(modA)## Family: Gamma ( log )
## Formula: biomass ~ U_landscape + U_local + (1 | plotid)
## Data: df_biomass
##
## AIC BIC logLik deviance df.resid
## 804.1 820.9 -395.1 790.1 74
##
## Random effects:
##
## Conditional model:
## Groups Name Variance Std.Dev.
## plotid (Intercept) 0.03642 0.1908
## Number of obs: 81, groups: plotid, 27
##
## Dispersion estimate for Gamma family (sigma^2): 0.42
##
## Conditional model:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 4.4694 0.1744 25.627 < 2e-16 ***
## U_landscapeMEDIUM -0.2351 0.2083 -1.129 0.259066
## U_landscapeHIGH -0.3265 0.1997 -1.635 0.102153
## U_localMEDIUM -0.2088 0.1893 -1.103 0.270124
## U_localHIGH -0.6096 0.1841 -3.312 0.000927 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Anova(modA, type="3") #U_local is significant## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: biomass
## Chisq Df Pr(>Chisq)
## (Intercept) 656.7630 1 < 2.2e-16 ***
## U_landscape 2.8112 2 0.245217
## U_local 11.5485 2 0.003107 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot(simulateResiduals(modA))
testDispersion(simulateResiduals(modA))
##
## DHARMa nonparametric dispersion test via sd of residuals fitted vs.
## simulated
##
## data: simulationOutput
## dispersion = 1.1348, p-value = 0.536
## alternative hypothesis: two.sidedplot(allEffects(modA))
contrast(emmeans(modA, specs=~U_local), method="pairwise") ## biomass different in low and high urbanised settings ## contrast estimate SE df z.ratio p.value
## LOW - MEDIUM 0.209 0.189 Inf 1.103 0.5123
## LOW - HIGH 0.610 0.184 Inf 3.312 0.0027
## MEDIUM - HIGH 0.401 0.181 Inf 2.209 0.0696
##
## Results are averaged over the levels of: U_landscape
## Results are given on the log (not the response) scale.
## P value adjustment: tukey method for comparing a family of 3 estimatespreds_modA <- emmeans(modA, specs=~U_local, type="response") |> as_tibble()
preds_modA## # A tibble: 3 × 6
## U_local response SE df asymp.LCL asymp.UCL
## <fct> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 LOW 72.4 9.55 Inf 55.9 93.8
## 2 MEDIUM 58.8 8.23 Inf 44.7 77.3
## 3 HIGH 39.4 5.31 Inf 30.2 51.3g_modA<-ggplot() +
geom_pointrange(data=preds_modA,
aes(x=U_local, y=response, ymin=asymp.LCL, ymax=asymp.UCL, fill=U_local),
position=position_dodge(width=0.5), size = 1, pch = 21)+
scale_fill_manual(values=urb.col3, name="")+
labs(x="urbanisation level at local scale",
y="biomass", ylim=c(18,26))+
ggtitle("effect of urbanisation at local scale on arthropod biomass")+
theme_bw()+
theme(axis.text = element_text(size=rel(1)),
axis.title = element_text(size=rel(1.1)),
plot.title = element_text(size=rel(1.1)))
g_modA
visualisation of statistics results of biomass:
g_modA_bis<-ggplot() +
geom_point(data=df_biomass, aes(x=U_local, y=biomass),
position=position_jitter(width = 0.2, height = 0), size=2, col="#B8C0C7")+
geom_pointrange(data=preds_modA,
aes(x=U_local, y=response, ymin=asymp.LCL, ymax=asymp.UCL, fill=U_local),
position=position_dodge(width=0.5), size = 1, pch = 21)+
scale_fill_manual(values=urb.col3, name="")+
labs(x="urbanisation level at local scale",
y=bquote('biomass'), ylim=c(18,26))+
ggtitle("effect of urbanisation level at local scale on arthropod biomass")+
theme_bw()+
theme(axis.text = element_text(size=rel(1)),
axis.title = element_text(size=rel(1.1)),
plot.title = element_text(size=rel(1.1)))
g_modA_bis
other type of visualisation:
ggplot(df_biomass, aes(U_landscape, biomass, colour= U_local)) +
geom_point(position=position_dodge(width=0.40))+
scale_colour_manual(values = urb.col3, name="urbanisation \nat local scale") +
ylab(bquote('biomass '(mm^3)))+
xlab("urbanisation level at landscape scale")+
theme_bw()+
ggtitle("effect of urbanisation on arthropod biomass")+
theme(axis.text = element_text(size=rel(1)), axis.title=element_text(size=rel(1)))
other type of visualisation:
b<-ggplot(df_biomass, aes(U_landscape, biomass, fill= U_local)) +
geom_boxplot(position=position_dodge(width=0.75))+
geom_jitter(position=position_dodge(width=0.75), alpha=0.6, size=1.5, color="gray3")+
scale_fill_manual(values = urb.col3, name="urbanisation \nat local scale") +
ylab(bquote('biomass '(mm^3)))+
xlab("urbanisation level at landscape scale")+
theme_bw()+
ggtitle("effect of urbanisation on arthropod biomass")+
theme(axis.text = element_text(size=rel(1)), axis.title=element_text(size=rel(1)))
b
prey size - average prey length
random factor : (1|plotid) fixed effects : urbanisation at landscape and local scale (interaction not significant)
response variable => log(length) to make it more normally distributed however levene test of residuals not ok
df <- data.st %>%
select(length, day) %>% ungroup()
df_cor<-cor(df, method = "pearson")
corrplot(df_cor,method = "circle", addgrid.col="grey", addCoef.col = "black", number.cex = 0.8, type="upper", title="correlation biomass and sampling period", mar=c(0,0,2,0))
#no correlation so we do not need to include day
modA<-glmmTMB(length ~ U_landscape+U_local+(1|plotid/location),data=data.st, family=gaussian)
summary(modA)## Family: gaussian ( identity )
## Formula: length ~ U_landscape + U_local + (1 | plotid/location)
## Data: data.st
##
## AIC BIC logLik deviance df.resid
## 36496.6 36553.9 -18240.3 36480.6 9597
##
## Random effects:
##
## Conditional model:
## Groups Name Variance Std.Dev.
## location:plotid (Intercept) 1.936e-01 0.4399602
## plotid (Intercept) 5.396e-07 0.0007346
## Residual 2.565e+00 1.6016842
## Number of obs: 9605, groups: location:plotid, 81; plotid, 27
##
## Dispersion estimate for gaussian family (sigma^2): 2.57
##
## Conditional model:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 2.47491 0.11590 21.353 <2e-16 ***
## U_landscapeMEDIUM -0.06789 0.12884 -0.527 0.5982
## U_landscapeHIGH -0.06869 0.12814 -0.536 0.5919
## U_localMEDIUM 0.04068 0.12853 0.316 0.7516
## U_localHIGH -0.21957 0.12840 -1.710 0.0873 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Anova(modA, type="3") ## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: length
## Chisq Df Pr(>Chisq)
## (Intercept) 455.9548 1 <2e-16 ***
## U_landscape 0.3785 2 0.8276
## U_local 4.6931 2 0.0957 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot(simulateResiduals(modA))
testDispersion(simulateResiduals(modA))
##
## DHARMa nonparametric dispersion test via sd of residuals fitted vs.
## simulated
##
## data: simulationOutput
## dispersion = 0.9994, p-value = 0.992
## alternative hypothesis: two.sidedtestZeroInflation(simulateResiduals(modA))
##
## DHARMa zero-inflation test via comparison to expected zeros with
## simulation under H0 = fitted model
##
## data: simulationOutput
## ratioObsSim = NaN, p-value = 1
## alternative hypothesis: two.sidedplot(allEffects(modA))
contrast(emmeans(modA, specs=~U_local), method="pairwise")## contrast estimate SE df t.ratio p.value
## LOW - MEDIUM -0.0407 0.129 9597 -0.316 0.9463
## LOW - HIGH 0.2196 0.128 9597 1.710 0.2014
## MEDIUM - HIGH 0.2602 0.129 9597 2.010 0.1099
##
## Results are averaged over the levels of: U_landscape
## P value adjustment: tukey method for comparing a family of 3 estimatesmodB<-glmmTMB(log(length) ~ U_landscape+U_local+(1|plotid/location),data=data.st, family=gaussian)
summary(modB)## Family: gaussian ( identity )
## Formula: log(length) ~ U_landscape + U_local + (1 | plotid/location)
## Data: data.st
##
## AIC BIC logLik deviance df.resid
## 16149.8 16207.2 -8066.9 16133.8 9597
##
## Random effects:
##
## Conditional model:
## Groups Name Variance Std.Dev.
## location:plotid (Intercept) 2.106e-02 1.451e-01
## plotid (Intercept) 3.105e-09 5.572e-05
## Residual 3.087e-01 5.556e-01
## Number of obs: 9605, groups: location:plotid, 81; plotid, 27
##
## Dispersion estimate for gaussian family (sigma^2): 0.309
##
## Conditional model:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.691153 0.038426 17.987 <2e-16 ***
## U_landscapeMEDIUM -0.020505 0.042781 -0.479 0.632
## U_landscapeHIGH 0.002699 0.042536 0.063 0.949
## U_localMEDIUM 0.001723 0.042664 0.040 0.968
## U_localHIGH -0.044774 0.042625 -1.050 0.294
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Anova(modB, type="3") ## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: log(length)
## Chisq Df Pr(>Chisq)
## (Intercept) 323.5202 1 <2e-16 ***
## U_landscape 0.3481 2 0.8403
## U_local 1.5111 2 0.4698
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot(simulateResiduals(modB))
testDispersion(simulateResiduals(modB))
##
## DHARMa nonparametric dispersion test via sd of residuals fitted vs.
## simulated
##
## data: simulationOutput
## dispersion = 1.0023, p-value = 0.872
## alternative hypothesis: two.sidedtestZeroInflation(simulateResiduals(modB))
##
## DHARMa zero-inflation test via comparison to expected zeros with
## simulation under H0 = fitted model
##
## data: simulationOutput
## ratioObsSim = Inf, p-value < 2.2e-16
## alternative hypothesis: two.sidedplot(allEffects(modB))
mod_test_gamma<-glmmTMB(length ~ U_landscape+U_local+(1|plotid/location),data=data.st, family=Gamma(link=log))
summary(mod_test_gamma)## Family: Gamma ( log )
## Formula: length ~ U_landscape + U_local + (1 | plotid/location)
## Data: data.st
##
## AIC BIC logLik deviance df.resid
## 30179.2 30236.6 -15081.6 30163.2 9597
##
## Random effects:
##
## Conditional model:
## Groups Name Variance Std.Dev.
## location:plotid (Intercept) 2.874e-02 0.1695297
## plotid (Intercept) 3.361e-08 0.0001833
## Number of obs: 9605, groups: location:plotid, 81; plotid, 27
##
## Dispersion estimate for Gamma family (sigma^2): 0.311
##
## Conditional model:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.903554 0.044203 20.441 <2e-16 ***
## U_landscapeMEDIUM -0.040398 0.049035 -0.824 0.4100
## U_landscapeHIGH -0.033503 0.048802 -0.687 0.4924
## U_localMEDIUM 0.002405 0.048930 0.049 0.9608
## U_localHIGH -0.100438 0.048887 -2.054 0.0399 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Anova(mod_test_gamma, type="3") ## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: length
## Chisq Df Pr(>Chisq)
## (Intercept) 417.8427 1 < 2e-16 ***
## U_landscape 0.7822 2 0.67631
## U_local 5.7096 2 0.05757 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot(simulateResiduals(mod_test_gamma))
testDispersion(simulateResiduals(mod_test_gamma))
##
## DHARMa nonparametric dispersion test via sd of residuals fitted vs.
## simulated
##
## data: simulationOutput
## dispersion = 1.3957, p-value < 2.2e-16
## alternative hypothesis: two.sidedtestZeroInflation(simulateResiduals(mod_test_gamma))
##
## DHARMa zero-inflation test via comparison to expected zeros with
## simulation under H0 = fitted model
##
## data: simulationOutput
## ratioObsSim = NaN, p-value = 1
## alternative hypothesis: two.sidedplot(allEffects(mod_test_gamma))
contrast(emmeans(mod_test_gamma, specs=~U_local), method="pairwise")## contrast estimate SE df z.ratio p.value
## LOW - MEDIUM -0.00241 0.0489 Inf -0.049 0.9987
## LOW - HIGH 0.10044 0.0489 Inf 2.054 0.0995
## MEDIUM - HIGH 0.10284 0.0493 Inf 2.088 0.0923
##
## Results are averaged over the levels of: U_landscape
## Results are given on the log (not the response) scale.
## P value adjustment: tukey method for comparing a family of 3 estimatesmod_test_pois<-glmmTMB(length ~ U_landscape+U_local+(1|plotid/location),data=data.st, family=poisson)## Warning in glmmTMB(length ~ U_landscape + U_local + (1 | plotid/location), :
## non-integer counts in a poisson modelsummary(mod_test_pois)## Family: poisson ( log )
## Formula: length ~ U_landscape + U_local + (1 | plotid/location)
## Data: data.st
##
## AIC BIC logLik deviance df.resid
## 33215.3 33265.5 -16600.6 33201.3 9598
##
## Random effects:
##
## Conditional model:
## Groups Name Variance Std.Dev.
## location:plotid (Intercept) 2.841e-02 0.1685532
## plotid (Intercept) 2.714e-08 0.0001648
## Number of obs: 9605, groups: location:plotid, 81; plotid, 27
##
## Conditional model:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.899999 0.044612 20.174 <2e-16 ***
## U_landscapeMEDIUM -0.037263 0.049806 -0.748 0.4544
## U_landscapeHIGH -0.031454 0.049473 -0.636 0.5249
## U_localMEDIUM 0.006068 0.049555 0.122 0.9025
## U_localHIGH -0.097941 0.049693 -1.971 0.0487 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Anova(mod_test_pois, type="3") #U_local is marginally significant## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: length
## Chisq Df Pr(>Chisq)
## (Intercept) 406.9794 1 < 2e-16 ***
## U_landscape 0.6550 2 0.72073
## U_local 5.4308 2 0.06618 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot(simulateResiduals(mod_test_pois))
testDispersion(simulateResiduals(mod_test_pois))
##
## DHARMa nonparametric dispersion test via sd of residuals fitted vs.
## simulated
##
## data: simulationOutput
## dispersion = 1.086, p-value = 0.024
## alternative hypothesis: two.sidedtestZeroInflation(simulateResiduals(mod_test_pois))
##
## DHARMa zero-inflation test via comparison to expected zeros with
## simulation under H0 = fitted model
##
## data: simulationOutput
## ratioObsSim = 0, p-value < 2.2e-16
## alternative hypothesis: two.sidedplot(allEffects(mod_test_pois))
mod_test_nbinom2<-glmmTMB(length ~ U_landscape+U_local+(1|plotid/location),data=data.st, family=poisson)## Warning in glmmTMB(length ~ U_landscape + U_local + (1 | plotid/location), :
## non-integer counts in a poisson modelsummary(mod_test_nbinom2)## Family: poisson ( log )
## Formula: length ~ U_landscape + U_local + (1 | plotid/location)
## Data: data.st
##
## AIC BIC logLik deviance df.resid
## 33215.3 33265.5 -16600.6 33201.3 9598
##
## Random effects:
##
## Conditional model:
## Groups Name Variance Std.Dev.
## location:plotid (Intercept) 2.841e-02 0.1685532
## plotid (Intercept) 2.714e-08 0.0001648
## Number of obs: 9605, groups: location:plotid, 81; plotid, 27
##
## Conditional model:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.899999 0.044612 20.174 <2e-16 ***
## U_landscapeMEDIUM -0.037263 0.049806 -0.748 0.4544
## U_landscapeHIGH -0.031454 0.049473 -0.636 0.5249
## U_localMEDIUM 0.006068 0.049555 0.122 0.9025
## U_localHIGH -0.097941 0.049693 -1.971 0.0487 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Anova(mod_test_nbinom2, type="3") #U_local is marginally significant## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: length
## Chisq Df Pr(>Chisq)
## (Intercept) 406.9794 1 < 2e-16 ***
## U_landscape 0.6550 2 0.72073
## U_local 5.4308 2 0.06618 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot(simulateResiduals(mod_test_nbinom2))
testDispersion(simulateResiduals(mod_test_nbinom2))
##
## DHARMa nonparametric dispersion test via sd of residuals fitted vs.
## simulated
##
## data: simulationOutput
## dispersion = 1.086, p-value = 0.024
## alternative hypothesis: two.sidedtestZeroInflation(simulateResiduals(mod_test_nbinom2))
##
## DHARMa zero-inflation test via comparison to expected zeros with
## simulation under H0 = fitted model
##
## data: simulationOutput
## ratioObsSim = 0, p-value < 2.2e-16
## alternative hypothesis: two.sidedplot(allEffects(mod_test_nbinom2))
statistics not yet fine: but gives that U_local is marginally significant (p = 0.06618) is probably the case => see visualisation :
### plot the distribution of the prey items
l1 <- data.st %>%
group_by(plotid, U_landscape) %>%
summarise(avg_length=mean(length, na.rm=T)) %>%
ggplot(aes(U_landscape, avg_length, fill= U_landscape)) +
geom_boxplot()+ geom_jitter(width=0.2)+
scale_fill_manual(values = urb.col3, guide="none") +
ylim(1.5,5)+
ylab("")+
ggtitle("landscape scale (3 km x 3 km)")+
theme_bw()+
theme(axis.title=element_text(size=rel(1.0)),
axis.text.x = element_text(size=rel(1.1)),
axis.text.y =element_text(size=rel(1.0)),
axis.title.x = element_blank())
l1 
l2 <- data.st %>%
group_by(plotid, U_local) %>%
summarise(avg_length=mean(length, na.rm=T)) %>%
ggplot(aes(U_local, avg_length, fill= U_local)) +
geom_boxplot()+ geom_jitter(width=0.2)+
scale_fill_manual(values = urb.col3, guide="none")+
ylim(1.5,5)+
ylab("Average prey length (mm)")+
ggtitle("local scale (200 m x 200 m)")+
theme_bw()+
theme(axis.title.y=element_text(size=rel(1.1)),
axis.text.x =element_text(size=rel(1.1)),
axis.text.y =element_text(size=rel(1)),
axis.title.x=element_blank())
l2
fig1<- ggarrange(l2, l1, ncol = 2)
fig1
annotate_figure(fig1, bottom=text_grob("Urbanisation level", vjust=0.5, size=14),
top=text_grob("Effect of urbanisation on average arthropod prey size", vjust=0.2, size=15))
### plot the distribution of the prey items
l1 <- data.st %>%
group_by(plotid, U_landscape) %>%
summarise(avg_n_prey=mean(sum(n_indiv))) %>%
ggplot(aes(U_landscape, avg_n_prey, fill= U_landscape)) +
geom_boxplot()+
geom_jitter(width=0.2)+
scale_fill_manual(values = urb.col3, guide="none") +
xlab("")+
ylab("")+
ylim(100,800)+
ggtitle("landscape scale (3 km x 3 km)")+
theme_bw()+
theme(axis.title=element_text(size=rel(1.2)),
axis.text.x = element_text(size=rel(1.2)),
axis.text.y =element_text(size=rel(1.2)),
axis.title.x = element_blank(),
axis.title.y= element_blank())
l1 
l2 <- data.st %>%
group_by(plotid, U_local) %>%
summarise(avg_n_prey=mean(sum(n_indiv))) %>%
ggplot(aes(U_local, avg_n_prey, fill= U_local)) +
geom_boxplot()+
geom_jitter(width=0.2)+
scale_fill_manual(values = urb.col3, guide="none") +
ylab("number of prey")+
ggtitle("local scale (200 m x 200 m)")+
theme_bw()+
theme(axis.title=element_text(size=rel(1.2)),
axis.text.x = element_text(size=rel(1.2)),
axis.text.y =element_text(size=rel(1.2)),
axis.title.x = element_blank())
l2
fig1<- ggarrange(l2, l1, ncol = 2)
fig1
annotate_figure(fig1, bottom=text_grob("Urbanization level", vjust=0.5, size=15))
prey diversity - nmds func spec
all functional species
func.matrix <- data.st %>%
group_by(location, plotid, U_landscape, U_local, urb_cat, day, sday, func_species) %>%
summarise(total_n_indiv=sum(n_indiv)) %>%
spread(key=func_species, value=total_n_indiv, fill=0) %>%
subset(select=-c(plotid, U_landscape, U_local, urb_cat, day, sday)) %>%
column_to_rownames(var="location")
##dit is de community matrix
info <- data.st %>%
group_by(location, plotid, U_landscape, U_local, urb_cat, day, sday, func_species) %>%
summarise(total_n_indiv=sum(n_indiv)) %>%
spread(key=func_species, value=total_n_indiv, fill=0) %>%
select(location, plotid, U_landscape, U_local, urb_cat, day, sday)
## dit is de info
info## # A tibble: 81 × 7
## # Groups: location, plotid, U_landscape, U_local, urb_cat, day, sday [81]
## location plotid U_landscape U_local urb_cat day sday[,1]
## <fct> <fct> <fct> <fct> <fct> <dbl> <dbl>
## 1 P01SG P01 HIGH LOW HIGHLOW 249 1.21
## 2 P01SY P01 HIGH MEDIUM HIGHMEDIUM 249 1.21
## 3 P01SR P01 HIGH HIGH HIGHHIGH 249 1.21
## 4 P02SG P02 HIGH LOW HIGHLOW 239 0.0798
## 5 P02SY P02 HIGH MEDIUM HIGHMEDIUM 239 0.0798
## 6 P02SR P02 HIGH HIGH HIGHHIGH 239 0.0798
## 7 P03SG P03 HIGH LOW HIGHLOW 229 -1.05
## 8 P03SY P03 HIGH MEDIUM HIGHMEDIUM 229 -1.05
## 9 P03SR P03 HIGH HIGH HIGHHIGH 229 -1.05
## 10 P04SG P04 MEDIUM LOW MEDIUMLOW 229 -1.05
## # ℹ 71 more rowshead(info)## # A tibble: 6 × 7
## # Groups: location, plotid, U_landscape, U_local, urb_cat, day, sday [6]
## location plotid U_landscape U_local urb_cat day sday[,1]
## <fct> <fct> <fct> <fct> <fct> <dbl> <dbl>
## 1 P01SG P01 HIGH LOW HIGHLOW 249 1.21
## 2 P01SY P01 HIGH MEDIUM HIGHMEDIUM 249 1.21
## 3 P01SR P01 HIGH HIGH HIGHHIGH 249 1.21
## 4 P02SG P02 HIGH LOW HIGHLOW 239 0.0798
## 5 P02SY P02 HIGH MEDIUM HIGHMEDIUM 239 0.0798
## 6 P02SR P02 HIGH HIGH HIGHHIGH 239 0.0798nmds_matrix=metaMDS(func.matrix, distance="bray", k=2, trymax=100, autotransform = F)## Run 0 stress 0.205799
## Run 1 stress 0.2068969
## Run 2 stress 0.21016
## Run 3 stress 0.2127645
## Run 4 stress 0.2145304
## Run 5 stress 0.2071227
## Run 6 stress 0.2254975
## Run 7 stress 0.2035478
## ... New best solution
## ... Procrustes: rmse 0.02687038 max resid 0.1680316
## Run 8 stress 0.2089785
## Run 9 stress 0.210042
## Run 10 stress 0.2088447
## Run 11 stress 0.2052119
## Run 12 stress 0.2104413
## Run 13 stress 0.2051957
## Run 14 stress 0.2100039
## Run 15 stress 0.2059824
## Run 16 stress 0.2028529
## ... New best solution
## ... Procrustes: rmse 0.01817394 max resid 0.1503437
## Run 17 stress 0.2478003
## Run 18 stress 0.20595
## Run 19 stress 0.2101974
## Run 20 stress 0.2354425
## Run 21 stress 0.2100393
## Run 22 stress 0.2082859
## Run 23 stress 0.211463
## Run 24 stress 0.2085858
## Run 25 stress 0.2068969
## Run 26 stress 0.2076207
## Run 27 stress 0.2102914
## Run 28 stress 0.2234853
## Run 29 stress 0.2111056
## Run 30 stress 0.2202592
## Run 31 stress 0.2151924
## Run 32 stress 0.234025
## Run 33 stress 0.2402233
## Run 34 stress 0.2368402
## Run 35 stress 0.2082478
## Run 36 stress 0.2204409
## Run 37 stress 0.2326045
## Run 38 stress 0.2185666
## Run 39 stress 0.2278931
## Run 40 stress 0.2028529
## ... New best solution
## ... Procrustes: rmse 0.000122142 max resid 0.0009709275
## ... Similar to previous best
## *** Best solution repeated 1 timesnmds_matrix##
## Call:
## metaMDS(comm = func.matrix, distance = "bray", k = 2, trymax = 100, autotransform = F)
##
## global Multidimensional Scaling using monoMDS
##
## Data: func.matrix
## Distance: bray
##
## Dimensions: 2
## Stress: 0.2028529
## Stress type 1, weak ties
## Best solution was repeated 1 time in 40 tries
## The best solution was from try 40 (random start)
## Scaling: centring, PC rotation, halfchange scaling
## Species: expanded scores based on 'func.matrix'sppr<- specnumber(func.matrix) #species richness
sppr## P01SG P01SY P01SR P02SG P02SY P02SR P03SG P03SY P03SR P04SG P04SY P04SR P05SG
## 29 28 28 38 33 20 24 35 31 25 33 31 57
## P05SY P05SR P06SG P06SY P06SR P07SG P07SY P07SR P08SG P08SY P08SR P09SG P09SY
## 33 25 45 26 23 46 42 28 49 36 22 56 36
## P09SR P10SG P10SY P10SR P11SG P11SY P11SR P12SG P12SY P12SR P13SG P13SY P13SR
## 29 53 20 29 49 31 28 40 46 26 37 11 27
## P14SG P14SY P14SR P15SG P15SY P15SR P16SG P16SY P16SR P17SG P17SY P17SR P18SG
## 43 39 41 30 54 24 26 34 32 37 44 42 44
## P18SY P18SR P19SG P19SY P19SR P20SG P20SY P20SR P21SG P21SY P21SR P22SG P22SY
## 27 28 32 26 20 16 28 30 26 26 36 23 22
## P22SR P23SG P23SY P23SR P24SG P24SY P24SR P25SG P25SY P25SR P26SG P26SY P26SR
## 32 31 33 19 13 17 21 27 23 26 35 36 34
## P27SG P27SY P27SR
## 26 23 28df_species_rich <- sppr %>%
enframe() %>%
full_join(info, by = c("name"="location")) %>%
rename("species_richness" = "value")
ggplot(df_species_rich, aes(x = U_landscape, y = species_richness, fill = U_local)) +
geom_boxplot(position=position_dodge(width=0.75)) +
geom_point(position=position_dodge(width=0.75), colour="gray3")+
scale_colour_manual(values = urb.col3) +
scale_fill_manual(values=urb.col3)+
theme_bw()+
#theme(legend.position = "none",
labs(x = "urbanisation level at local scale",
y = "Number of functional species per subplot",
title = "Species richness")
number of species decreases with increasing urbanisation at local
scale
###statistics species richness species richness : started with poisson but overdispersion so modified to nbinm2 => U_local significant
df <- df_species_rich %>%
select(species_richness, day) %>% ungroup()
df_cor<-cor(df, method = "pearson")
corrplot(df_cor,method = "circle", addgrid.col="grey", addCoef.col = "black", number.cex = 0.8, type="upper", title="correlation species richness and sampling day", mar=c(0,0,2,0))
modA<-glmmTMB(species_richness ~ U_landscape+U_local+(1|plotid), data=df_species_rich, family=nbinom2) #nbinom 2 to correct for overdispersion
summary(modA)## Family: nbinom2 ( log )
## Formula: species_richness ~ U_landscape + U_local + (1 | plotid)
## Data: df_species_rich
##
## AIC BIC logLik deviance df.resid
## 594.0 610.8 -290.0 580.0 74
##
## Random effects:
##
## Conditional model:
## Groups Name Variance Std.Dev.
## plotid (Intercept) 0.00804 0.08966
## Number of obs: 81, groups: plotid, 27
##
## Dispersion parameter for nbinom2 family (): 25
##
## Conditional model:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 3.63379 0.07171 50.67 < 2e-16 ***
## U_landscapeMEDIUM -0.11987 0.08435 -1.42 0.15528
## U_landscapeHIGH -0.09885 0.08411 -1.18 0.23991
## U_localMEDIUM -0.12711 0.07222 -1.76 0.07839 .
## U_localHIGH -0.22610 0.07320 -3.09 0.00201 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Anova(modA, type="3") #U_local is significant## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: species_richness
## Chisq Df Pr(>Chisq)
## (Intercept) 2567.5276 1 < 2e-16 ***
## U_landscape 2.3217 2 0.31323
## U_local 9.6391 2 0.00807 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot(simulateResiduals(modA))
testDispersion(simulateResiduals(modA))
##
## DHARMa nonparametric dispersion test via sd of residuals fitted vs.
## simulated
##
## data: simulationOutput
## dispersion = 1.0356, p-value = 0.768
## alternative hypothesis: two.sidedplot(allEffects(modA))## Warning in Effect.glmmTMB(predictors, mod, vcov. = vcov., ...): overriding
## variance function for effects/dev.resids: computed variances may be incorrect
## Warning in Effect.glmmTMB(predictors, mod, vcov. = vcov., ...): overriding
## variance function for effects/dev.resids: computed variances may be incorrect
contrast(emmeans(modA, specs=~U_local), method="pairwise") ## contrast estimate SE df z.ratio p.value
## LOW - MEDIUM 0.127 0.0722 Inf 1.760 0.1832
## LOW - HIGH 0.226 0.0732 Inf 3.089 0.0057
## MEDIUM - HIGH 0.099 0.0741 Inf 1.336 0.3754
##
## Results are averaged over the levels of: U_landscape
## Results are given on the log (not the response) scale.
## P value adjustment: tukey method for comparing a family of 3 estimatespreds_modA <- emmeans(modA, specs=~U_local, type="response") |> as_tibble()
preds_modA## # A tibble: 3 × 6
## U_local response SE df asymp.LCL asymp.UCL
## <fct> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 LOW 35.2 1.89 Inf 31.7 39.1
## 2 MEDIUM 31.0 1.70 Inf 27.8 34.5
## 3 HIGH 28.1 1.56 Inf 25.2 31.3g_sp<-ggplot()+
geom_point(data=df_species_rich, aes(x=U_local, y=species_richness),
position=position_jitter(width = 0.2, height = 0), size=2, col="#B8C0C7")+
geom_pointrange(data=preds_modA,
aes(x=U_local, y=response, ymin=asymp.LCL, ymax=asymp.UCL, fill=U_local), position=position_dodge(width=0.5), size = 1, pch = 21)+
scale_fill_manual(values=urb.col3, name="")+
labs(x="urbanisation level at local scale", y="species richness")+
ggtitle("effect of urbanisation on species richness")+
theme_bw()
g_sp
NMDS
## new way:
NMDS.vect <- scores(nmds_matrix, display = "species") %>%
as.data.frame()
NMDS.scores <- scores(nmds_matrix, display = "sites") %>%
as.data.frame()%>%
rownames_to_column("location") %>%
full_join(info, by="location")
stressplot(nmds_matrix)
##define a theme to use later
clean_background <- theme(plot.background = element_rect("white"),
panel.background = element_rect("white"),
panel.grid = element_line("white"),
axis.line = element_line("gray25"),
axis.text = element_text(size = 12, color = "gray25"),
axis.title = element_text(color = "gray25"),
legend.text = element_text(size = 12),
legend.key = element_rect("white"))
ordbray.scores <- as.data.frame(nmds_matrix$points) %>%
bind_cols(info)
ordbray.scores$U_local <- factor(ordbray.scores$U_local, levels=c("LOW", "MEDIUM", "HIGH"))
ordbray.scores$U_landscape <- factor(ordbray.scores$U_landscape, levels=c("LOW", "MEDIUM", "HIGH"))
ordbray.hulls <- ordbray.scores %>%
group_by(U_local) %>%
slice(chull(MDS1, MDS2))
# plot nMDS with color and convex hulls as U_local
ordbray.plot.mds <- ggplot() +
geom_point(data = ordbray.scores, aes(x = MDS1, y = MDS2, color = U_local), size = 2) +
geom_polygon(data = ordbray.hulls, aes(x = MDS1, MDS2, fill = U_local), alpha = 0.5) +
scale_fill_manual(values=urb.col3) +
scale_colour_manual(values=urb.col3)+
theme_bw() +
xlab("NMDS1") +
ylab("NMDS2") +
ggtitle("ordination functional species : nmds bray curtis dissimilarity ")
ordbray.plot.mds
envfit on func species
# envfit() takes the output of metaMDS() and the species matrix you created
fit <- envfit(nmds_matrix, func.matrix, perm = 999)
#extract p-values for each species
fit_pvals <- fit$vectors$pvals %>%
as.data.frame() %>%
rownames_to_column("func_species") %>%
dplyr::rename("pvals" = ".")
# extract coordinates for species, only keep species with p-val = 0.001
fit_spp <- fit %>%
scores(., display = "vectors") %>%
as.data.frame() %>%
rownames_to_column("func_species") %>%
full_join(., fit_pvals, by = "func_species") %>%
filter(pvals <= 0.001)
set.seed(4)
nmds_plot_new <- ggplot(NMDS.scores, aes(x = NMDS1, y = NMDS2)) +
coord_fixed() +
geom_point(aes(color = U_local, shape = U_landscape), size = 3, alpha = 0.8) +
stat_ellipse(aes(color = U_local)) +
scale_colour_manual(name= "Urbanisation \nat local scale", values= urb.col3)+
scale_shape_manual(name= "urbanisation level \nat landscape scale", values=c(16, 17, 15))+
geom_segment(data = fit_spp, aes(x = 0, xend = NMDS1, y = 0, yend = NMDS2),
arrow = arrow(length = unit(0.25, "cm")),
col = "black") +
geom_text(data = fit_spp, aes(label = func_species), position = position_jitter(width = 0.3, height=0.1)) +
clean_background
nmds_plot_new
##permanova full dataset
library(pairwiseAdonis)
ST.dist<-vegdist(func.matrix, distance="bray")
perman <- adonis2(ST.dist~ plotid + U_landscape * U_local,
data=info,
strata=info$plotid,
permutations=999)
perman## Permutation test for adonis under reduced model
## Terms added sequentially (first to last)
## Blocks: strata
## Permutation: free
## Number of permutations: 999
##
## adonis2(formula = ST.dist ~ plotid + U_landscape * U_local, data = info, permutations = 999, strata = info$plotid)
## Df SumOfSqs R2 F Pr(>F)
## plotid 26 7.3473 0.40003 1.4122 0.09 .
## U_local 2 0.6025 0.03280 1.5054 0.03 *
## U_landscape:U_local 4 0.8118 0.04420 1.0142 0.44
## Residual 48 9.6052 0.52297
## Total 80 18.3668 1.00000
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1knitr::kable(perman)| Df | SumOfSqs | R2 | F | Pr(>F) | |
|---|---|---|---|---|---|
| plotid | 26 | 7.3472568 | 0.4000300 | 1.412166 | 0.09 |
| U_local | 2 | 0.6024680 | 0.0328021 | 1.505352 | 0.03 |
| U_landscape:U_local | 4 | 0.8118197 | 0.0442005 | 1.014223 | 0.44 |
| Residual | 48 | 9.6052182 | 0.5229674 | NA | NA |
| Total | 80 | 18.3667627 | 1.0000000 | NA | NA |
bd_local <- betadisper(ST.dist, info$U_local)
bd_local ##
## Homogeneity of multivariate dispersions
##
## Call: betadisper(d = ST.dist, group = info$U_local)
##
## No. of Positive Eigenvalues: 58
## No. of Negative Eigenvalues: 22
##
## Average distance to median:
## LOW MEDIUM HIGH
## 0.4536 0.4498 0.4743
##
## Eigenvalues for PCoA axes:
## (Showing 8 of 80 eigenvalues)
## PCoA1 PCoA2 PCoA3 PCoA4 PCoA5 PCoA6 PCoA7 PCoA8
## 2.8477 2.0489 1.1139 0.8832 0.8674 0.8456 0.6520 0.6361print(permutest(bd_local,pairwise=TRUE,permutations=999)) ## als het niet significant is, dan is het goed##
## Permutation test for homogeneity of multivariate dispersions
## Permutation: free
## Number of permutations: 999
##
## Response: Distances
## Df Sum Sq Mean Sq F N.Perm Pr(>F)
## Groups 2 0.00941 0.0047038 0.4998 999 0.615
## Residuals 78 0.73402 0.0094105
##
## Pairwise comparisons:
## (Observed p-value below diagonal, permuted p-value above diagonal)
## LOW MEDIUM HIGH
## LOW 0.89100 0.460
## MEDIUM 0.88872 0.318
## HIGH 0.45456 0.32280bd_land <- betadisper(ST.dist, info$U_landscape)
print(permutest(bd_land,pairwise=TRUE,permutations=999))##
## Permutation test for homogeneity of multivariate dispersions
## Permutation: free
## Number of permutations: 999
##
## Response: Distances
## Df Sum Sq Mean Sq F N.Perm Pr(>F)
## Groups 2 0.02302 0.0115095 1.4515 999 0.226
## Residuals 78 0.61848 0.0079292
##
## Pairwise comparisons:
## (Observed p-value below diagonal, permuted p-value above diagonal)
## LOW MEDIUM HIGH
## LOW 0.242000 0.070
## MEDIUM 0.254571 0.675
## HIGH 0.075275 0.691879no signs of “dispersion” => betadisper test is ok
info$interaction_comb <- interaction(info$U_landscape, info$U_local)
info$interaction_comb <- as.factor(info$interaction_comb)
pairwise.adonis(ST.dist, info$U_local) |> as_tibble() |> knitr::kable()| pairs | Df | SumsOfSqs | F.Model | R2 | p.value | p.adjusted | sig |
|---|---|---|---|---|---|---|---|
| LOW vs MEDIUM | 1 | 0.3389884 | 1.5344068 | 0.0286621 | 0.069 | 0.207 | |
| LOW vs HIGH | 1 | 0.3448402 | 1.4804737 | 0.0276825 | 0.085 | 0.255 | |
| MEDIUM vs HIGH | 1 | 0.2198734 | 0.9585059 | 0.0180992 | 0.509 | 1.000 |
pairwise.adonis(ST.dist, info$U_landscape)## pairs Df SumsOfSqs F.Model R2 p.value p.adjusted sig
## 1 HIGH vs MEDIUM 1 0.1638342 0.6817194 0.01294034 0.882 1.000
## 2 HIGH vs LOW 1 0.3235652 1.4408247 0.02696112 0.078 0.234
## 3 MEDIUM vs LOW 1 0.2672431 1.2028779 0.02260926 0.197 0.591no significant difference in communities
focus on top 3 orders
data.st.dhh <- data.st %>%
dplyr::filter(order %in% top3 & location !="P24SG" & location !="P20SG") %>%
droplevels()
dhh.matrix <- data.st.dhh %>%
dplyr::group_by(location, plotid, U_landscape, U_local, urb_cat, func_species) %>%
dplyr::summarise(total_n_indiv=sum(n_indiv)) %>%
spread(key=func_species, value=total_n_indiv, fill=0) %>%
subset(select=-c(plotid, U_landscape, U_local, urb_cat)) %>%
column_to_rownames(var="location")
##dit is de community matrix
##sum(func.matrix) gives now 8812
info <- data.st.dhh %>%
dplyr::group_by(location, plotid, U_landscape, U_local,urb_cat, func_species) %>%
dplyr::summarise(total_n_indiv=sum(n_indiv)) %>%
spread(key=func_species, value=total_n_indiv, fill=0) %>%
select(location, plotid, U_landscape, U_local, urb_cat)
## dit is de info
#func.hell <- decostand(dhh.matrix, method = "hellinger") #veranderde niets
nmds_matrix=metaMDS(dhh.matrix, distance="bray", k=2)## Square root transformation
## Wisconsin double standardization
## Run 0 stress 0.2931411
## Run 1 stress 0.2948932
## Run 2 stress 0.3022437
## Run 3 stress 0.3072661
## Run 4 stress 0.296218
## Run 5 stress 0.3004433
## Run 6 stress 0.3043133
## Run 7 stress 0.3032487
## Run 8 stress 0.3058203
## Run 9 stress 0.3093865
## Run 10 stress 0.2934998
## ... Procrustes: rmse 0.0285168 max resid 0.1294497
## Run 11 stress 0.3072166
## Run 12 stress 0.2986092
## Run 13 stress 0.3067337
## Run 14 stress 0.296813
## Run 15 stress 0.3143007
## Run 16 stress 0.3055637
## Run 17 stress 0.3064587
## Run 18 stress 0.2978105
## Run 19 stress 0.2964639
## Run 20 stress 0.2977014
## *** Best solution was not repeated -- monoMDS stopping criteria:
## 5: no. of iterations >= maxit
## 15: stress ratio > sratmaxnmds_matrix##
## Call:
## metaMDS(comm = dhh.matrix, distance = "bray", k = 2)
##
## global Multidimensional Scaling using monoMDS
##
## Data: wisconsin(sqrt(dhh.matrix))
## Distance: bray
##
## Dimensions: 2
## Stress: 0.2931411
## Stress type 1, weak ties
## Best solution was not repeated after 20 tries
## The best solution was from try 0 (metric scaling or null solution)
## Scaling: centring, PC rotation, halfchange scaling
## Species: expanded scores based on 'wisconsin(sqrt(dhh.matrix))'sppr<- specnumber(dhh.matrix)
sppr## P01SG P01SY P01SR P02SG P02SY P02SR P03SG P03SY P03SR P04SG P04SY P04SR P05SG
## 21 19 23 30 25 19 16 32 25 22 26 28 45
## P05SY P05SR P06SG P06SY P06SR P07SG P07SY P07SR P08SG P08SY P08SR P09SG P09SY
## 27 20 40 22 17 37 37 24 37 27 18 44 28
## P09SR P10SG P10SY P10SR P11SG P11SY P11SR P12SG P12SY P12SR P13SG P13SY P13SR
## 22 44 17 21 40 27 24 34 37 23 31 11 24
## P14SG P14SY P14SR P15SG P15SY P15SR P16SG P16SY P16SR P17SG P17SY P17SR P18SG
## 36 31 34 26 43 22 24 31 29 31 34 34 36
## P18SY P18SR P19SG P19SY P19SR P20SY P20SR P21SG P21SY P21SR P22SG P22SY P22SR
## 23 25 29 20 17 25 28 21 20 32 20 20 26
## P23SG P23SY P23SR P24SY P24SR P25SG P25SY P25SR P26SG P26SY P26SR P27SG P27SY
## 28 27 15 14 16 25 20 18 29 25 28 23 18
## P27SR
## 26df_species_rich <- sppr %>%
enframe() %>%
full_join(info, by = c("name"="location")) %>%
rename("species_richness" = "value")
#modA<-glmmTMB(species_richness ~ U_landscape+U_local+(1|plotid), data=df_species_rich, family=nbinom2) #nbinom 2 to correct for overdispersion
#summary(modA)
#Anova(modA, type="3") #U_local is significant
#plot(simulateResiduals(modA))
#testDispersion(simulateResiduals(modA))
#plot(allEffects(modA))
top3_rich<-ggplot(df_species_rich, aes(x = U_landscape, y = species_richness, fill = U_local)) +
geom_boxplot(position=position_dodge(width=0.75)) +
geom_point(position=position_dodge(width=0.75), colour="gray3")+
scale_colour_manual(values = urb.col3) +
scale_fill_manual(values=urb.col3)+
theme_bw()+
#theme(legend.position = "none",
labs(x = "urbanisation level at local scale",
y = "Number of functional species per subplot",
title = "Species richness funct species top 3 orders")
#top3_rich ## silence for now (give all the same trend as the full dataset)
## new way:
NMDS.vect <- scores(nmds_matrix, display = "species") %>%
as.data.frame()
NMDS.scores <- scores(nmds_matrix, display = "sites") %>%
as.data.frame()%>%
rownames_to_column("location") %>%
full_join(info, by="location")
stressplot(nmds_matrix)
# envfit() takes the output of metaMDS() and the species matrix you created
fit <- envfit(nmds_matrix, dhh.matrix, perm = 999)
#extract p-values for each species
fit_pvals <- fit$vectors$pvals %>%
as.data.frame() %>%
rownames_to_column("func_species") %>%
dplyr::rename("pvals" = ".")
# extract coordinates for species, only keep species with p-val = 0.001
fit_spp <- fit %>%
scores(., display = "vectors") %>%
as.data.frame() %>%
rownames_to_column("func_species") %>%
full_join(., fit_pvals, by = "func_species") %>%
filter(pvals <= 0.001)
set.seed(4)
nmds_plot_new <- ggplot(NMDS.scores, aes(x = NMDS1, y = NMDS2)) +
coord_fixed() +
geom_point(aes(color = U_local, shape = U_landscape), size = 3, alpha = 0.8) +
stat_ellipse(aes(color = U_local)) +
scale_color_manual(name="urbanisation level \nat local scale", values = urb.col3) +
scale_shape_manual(name= "urbanisation level \nat landscape scale", values=c(16, 17, 15))+
geom_segment(data = fit_spp, aes(x = 0, xend = NMDS1, y = 0, yend = NMDS2),
arrow = arrow(length = unit(0.25, "cm")),
col = "black") +
geom_text(data = fit_spp, aes(label = func_species), position = position_jitter(width = 0.3, height=0.1)) +
clean_background
nmds_plot_new
ordbray.scores <- as.data.frame(nmds_matrix$points) %>%
bind_cols(info)
ordbray.scores$U_local <- factor(ordbray.scores$U_local, levels=c("LOW", "MEDIUM", "HIGH"))
ordbray.scores$U_landscape <- factor(ordbray.scores$U_landscape, levels=c("LOW", "MEDIUM", "HIGH"))
ordbray.hulls <- ordbray.scores %>%
group_by(U_local) %>%
slice(chull(MDS1, MDS2))
# plot nMDS with color and convex hulls as U_local
ordbray.plot.mds <- ggplot() +
geom_point(data = ordbray.scores, aes(x = MDS1, y = MDS2, color = U_local), size = 2) +
geom_polygon(data = ordbray.hulls, aes(x = MDS1, MDS2, fill = U_local), alpha = 0.5) +
scale_fill_manual(values=urb.col3) +
scale_colour_manual(values=urb.col3)+
theme_bw() +
xlab("NMDS1") +
ylab("NMDS2") +
ggtitle("ordination top3 order functional species : nmds bray curtis dissimilarity ")
ordbray.plot.mds
ordbray.scores <- as.data.frame(nmds_matrix$points) %>%
bind_cols(info)
ordbray.scores$U_local <- factor(ordbray.scores$U_local, levels=c("LOW", "MEDIUM", "HIGH"))
ordbray.scores$U_landscape <- factor(ordbray.scores$U_landscape, levels=c("LOW", "MEDIUM", "HIGH"))
ordbray.hulls <- ordbray.scores %>%
group_by(urb_cat) %>%
slice(chull(MDS1, MDS2))
# plot nMDS with color and convex hulls as U_local
ordbray.plot.mds <- ggplot() +
geom_point(data = ordbray.scores, aes(x = MDS1, y = MDS2, color = urb_cat), size = 2) +
geom_polygon(data = ordbray.hulls, aes(x = MDS1, MDS2, fill = urb_cat), alpha = 0.5) +
scale_fill_manual(values=urb.col9) +
scale_colour_manual(values=urb.col9)+
theme_bw() +
xlab("NMDS1") +
ylab("NMDS2") +
ggtitle("ordination top3 order functional species : nmds bray curtis dissimilarity ")
ordbray.plot.mds
top 3 orders permanova
library(pairwiseAdonis)
ST.dist<-vegdist(dhh.matrix, distance="bray")
perman <- adonis2(ST.dist~ plotid + U_landscape * U_local,
data=info,
strata=info$plotid,
permutations=999)
perman## Permutation test for adonis under reduced model
## Terms added sequentially (first to last)
## Blocks: strata
## Permutation: free
## Number of permutations: 999
##
## adonis2(formula = ST.dist ~ plotid + U_landscape * U_local, data = info, permutations = 999, strata = info$plotid)
## Df SumOfSqs R2 F Pr(>F)
## plotid 26 6.8955 0.41449 1.4555 0.054 .
## U_local 2 0.6517 0.03918 1.7885 0.008 **
## U_landscape:U_local 4 0.7075 0.04253 0.9707 0.571
## Residual 46 8.3815 0.50381
## Total 78 16.6362 1.00000
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1knitr::kable(perman)| Df | SumOfSqs | R2 | F | Pr(>F) | |
|---|---|---|---|---|---|
| plotid | 26 | 6.8954858 | 0.4144864 | 1.4555461 | 0.054 |
| U_local | 2 | 0.6517387 | 0.0391759 | 1.7884547 | 0.008 |
| U_landscape:U_local | 4 | 0.7074600 | 0.0425253 | 0.9706806 | 0.571 |
| Residual | 46 | 8.3815314 | 0.5038124 | NA | NA |
| Total | 78 | 16.6362158 | 1.0000000 | NA | NA |
bd_local <- betadisper(ST.dist, info$U_local)
bd_local ##
## Homogeneity of multivariate dispersions
##
## Call: betadisper(d = ST.dist, group = info$U_local)
##
## No. of Positive Eigenvalues: 54
## No. of Negative Eigenvalues: 24
##
## Average distance to median:
## LOW MEDIUM HIGH
## 0.4172 0.4391 0.4631
##
## Eigenvalues for PCoA axes:
## (Showing 8 of 78 eigenvalues)
## PCoA1 PCoA2 PCoA3 PCoA4 PCoA5 PCoA6 PCoA7 PCoA8
## 2.9265 2.0659 1.1090 0.9075 0.8422 0.7317 0.6595 0.6253print(permutest(bd_local,pairwise=TRUE,permutations=999)) ## als het niet significant is, dan is het goed##
## Permutation test for homogeneity of multivariate dispersions
## Permutation: free
## Number of permutations: 999
##
## Response: Distances
## Df Sum Sq Mean Sq F N.Perm Pr(>F)
## Groups 2 0.02748 0.0137417 1.5098 999 0.229
## Residuals 76 0.69172 0.0091016
##
## Pairwise comparisons:
## (Observed p-value below diagonal, permuted p-value above diagonal)
## LOW MEDIUM HIGH
## LOW 0.417000 0.098
## MEDIUM 0.400740 0.365
## HIGH 0.094216 0.361372bd_land <- betadisper(ST.dist, info$U_landscape)
print(permutest(bd_land,pairwise=TRUE,permutations=999))##
## Permutation test for homogeneity of multivariate dispersions
## Permutation: free
## Number of permutations: 999
##
## Response: Distances
## Df Sum Sq Mean Sq F N.Perm Pr(>F)
## Groups 2 0.01743 0.0087134 1.0954 999 0.37
## Residuals 76 0.60452 0.0079542
##
## Pairwise comparisons:
## (Observed p-value below diagonal, permuted p-value above diagonal)
## LOW MEDIUM HIGH
## LOW 0.54500 0.119
## MEDIUM 0.52813 0.460
## HIGH 0.12164 0.42801no signs of “dispersion” => betadisper test is ok
info$interaction_comb <- interaction(info$U_landscape, info$U_local)
info$interaction_comb <- as.factor(info$interaction_comb)
pairwise.adonis(ST.dist, info$U_local) |> as_tibble() |> knitr::kable()| pairs | Df | SumsOfSqs | F.Model | R2 | p.value | p.adjusted | sig |
|---|---|---|---|---|---|---|---|
| LOW vs MEDIUM | 1 | 0.3556410 | 1.7858194 | 0.0344847 | 0.047 | 0.141 | |
| LOW vs HIGH | 1 | 0.4084134 | 1.9312649 | 0.0371889 | 0.023 | 0.069 | |
| MEDIUM vs HIGH | 1 | 0.2166091 | 0.9846406 | 0.0185835 | 0.468 | 1.000 |
pairwise.adonis(ST.dist, info$U_landscape)## pairs Df SumsOfSqs F.Model R2 p.value p.adjusted sig
## 1 HIGH vs MEDIUM 1 0.1397605 0.6311697 0.01246603 0.901 1.000
## 2 HIGH vs LOW 1 0.2979443 1.4014190 0.02674391 0.119 0.357
## 3 MEDIUM vs LOW 1 0.2287097 1.1143197 0.02138222 0.305 0.915focus on diptera
data.st.dip <- data.st %>%
dplyr::filter(order == "DIPTERA") %>%
droplevels()
dip.matrix <- data.st.dip %>%
dplyr::group_by(location, plotid, U_landscape, U_local, urb_cat, func_species) %>%
dplyr::summarise(total_n_indiv=sum(n_indiv)) %>%
spread(key=func_species, value=total_n_indiv, fill=0) %>%
subset(select=-c(plotid, U_landscape, U_local, urb_cat)) %>%
column_to_rownames(var="location")
##dit is de community matrix
##sum(dip.matrix) gives now 5287
info <- data.st.dip %>%
dplyr::group_by(location, plotid, U_landscape, U_local, urb_cat, func_species) %>%
dplyr::summarise(total_n_indiv=sum(n_indiv)) %>%
spread(key=func_species, value=total_n_indiv, fill=0) %>%
select(location, plotid, U_landscape, U_local, urb_cat)
## dit is de info
nmds_matrix=metaMDS(dip.matrix, distance="bray")## Square root transformation
## Wisconsin double standardization
## Run 0 stress 0.2647948
## Run 1 stress 0.2802532
## Run 2 stress 0.2759063
## Run 3 stress 0.2692734
## Run 4 stress 0.2718641
## Run 5 stress 0.2776076
## Run 6 stress 0.2840558
## Run 7 stress 0.2647879
## ... New best solution
## ... Procrustes: rmse 0.003919021 max resid 0.02634658
## Run 8 stress 0.2814574
## Run 9 stress 0.2793722
## Run 10 stress 0.2698903
## Run 11 stress 0.2811248
## Run 12 stress 0.2693693
## Run 13 stress 0.2758939
## Run 14 stress 0.2822895
## Run 15 stress 0.2786681
## Run 16 stress 0.2687636
## Run 17 stress 0.2794938
## Run 18 stress 0.2766532
## Run 19 stress 0.2751782
## Run 20 stress 0.282888
## *** Best solution was not repeated -- monoMDS stopping criteria:
## 4: no. of iterations >= maxit
## 16: stress ratio > sratmaxnmds_matrix##
## Call:
## metaMDS(comm = dip.matrix, distance = "bray")
##
## global Multidimensional Scaling using monoMDS
##
## Data: wisconsin(sqrt(dip.matrix))
## Distance: bray
##
## Dimensions: 2
## Stress: 0.2647879
## Stress type 1, weak ties
## Best solution was not repeated after 20 tries
## The best solution was from try 7 (random start)
## Scaling: centring, PC rotation, halfchange scaling
## Species: expanded scores based on 'wisconsin(sqrt(dip.matrix))'sppr<- specnumber(dip.matrix)
sppr## P01SG P01SY P01SR P02SG P02SY P02SR P03SG P03SY P03SR P04SG P04SY P04SR P05SG
## 13 10 10 19 13 10 9 20 14 14 10 18 26
## P05SY P05SR P06SG P06SY P06SR P07SG P07SY P07SR P08SG P08SY P08SR P09SG P09SY
## 12 10 22 13 10 22 19 12 21 18 7 25 17
## P09SR P10SG P10SY P10SR P11SG P11SY P11SR P12SG P12SY P12SR P13SG P13SY P13SR
## 13 20 7 8 20 17 12 18 18 12 15 5 8
## P14SG P14SY P14SR P15SG P15SY P15SR P16SG P16SY P16SR P17SG P17SY P17SR P18SG
## 18 18 20 15 22 14 12 16 14 19 21 19 21
## P18SY P18SR P19SG P19SY P19SR P20SG P20SY P20SR P21SG P21SY P21SR P22SG P22SY
## 13 12 17 9 7 7 14 13 8 11 13 12 11
## P22SR P23SG P23SY P23SR P24SG P24SY P24SR P25SG P25SY P25SR P26SG P26SY P26SR
## 8 13 15 3 7 9 8 15 11 10 17 16 17
## P27SG P27SY P27SR
## 13 9 14df_species_rich <- sppr %>%
enframe() %>%
full_join(info, by = c("name"="location")) %>%
rename("species_richness" = "value")
#modA<-glmmTMB(species_richness ~ U_landscape+U_local+(1|plotid), data=df_species_rich, family=poisson) #poisson (no overdispersion)
#summary(modA)
#Anova(modA, type="3") #U_local is significant
#plot(simulateResiduals(modA))
#testDispersion(simulateResiduals(modA))
#plot(allEffects(modA))
dip_sp<-ggplot(df_species_rich, aes(x = U_landscape, y = species_richness, fill = U_local)) +
geom_boxplot(position=position_dodge(width=0.75)) +
geom_point(position=position_dodge(width=0.75), colour="gray3")+
scale_colour_manual(values = urb.col3) +
scale_fill_manual(values=urb.col3)+
theme_bw()+
#theme(legend.position = "none",
labs(x = "urbanisation level at local scale",
y = "Number of functional species per subplot",
title = "Species richness funct species DIPTERA")
#dip_sp
### graph resembles previous because those orders make up the majority of the species
## new way:
NMDS.vect <- scores(nmds_matrix, display = "species") %>%
as.data.frame()
NMDS.scores <- scores(nmds_matrix, display = "sites") %>%
as.data.frame()%>%
rownames_to_column("location") %>%
full_join(info, by="location")
stressplot(nmds_matrix)
# envfit() takes the output of metaMDS() and the species matrix you created
fit <- envfit(nmds_matrix, dip.matrix, perm = 999)
#extract p-values for each species
fit_pvals <- fit$vectors$pvals %>%
as.data.frame() %>%
rownames_to_column("func_species") %>%
dplyr::rename("pvals" = ".")
# extract coordinates for species, only keep species with p-val = 0.001
fit_spp <- fit %>%
scores(., display = "vectors") %>%
as.data.frame() %>%
rownames_to_column("func_species") %>%
full_join(., fit_pvals, by = "func_species") %>%
filter(pvals <= 0.001)
nmds_plot_new <- ggplot(NMDS.scores, aes(x = NMDS1, y = NMDS2)) +
coord_fixed() +
geom_point(aes(color = U_local, shape = U_landscape), size = 3, alpha = 0.8) +
stat_ellipse(aes(color = U_local)) +
scale_colour_manual(name= "urbanisation level \nat local scale", values = urb.col3)+
scale_shape_manual(name= "urbanisation level \nat landscape scale", values=c(16, 17, 15))+
geom_segment(data = fit_spp, aes(x = 0, xend = NMDS1, y = 0, yend = NMDS2),
arrow = arrow(length = unit(0.25, "cm")),
col = "black") +
geom_text(data = fit_spp, aes(label = func_species), position = position_jitter(width = 0.2, height=0.1)) +
clean_background
nmds_plot_new
###permanova diptera
library(pairwiseAdonis)
ST.dist<-vegdist(dip.matrix, distance="bray")
perman <- adonis2(ST.dist~ plotid + U_landscape * U_local,
data=info,
strata=info$plotid,
permutations=999)
perman## Permutation test for adonis under reduced model
## Terms added sequentially (first to last)
## Blocks: strata
## Permutation: free
## Number of permutations: 999
##
## adonis2(formula = ST.dist ~ plotid + U_landscape * U_local, data = info, permutations = 999, strata = info$plotid)
## Df SumOfSqs R2 F Pr(>F)
## plotid 26 5.9884 0.39988 1.4151 0.142
## U_local 2 0.6420 0.04287 1.9721 0.018 *
## U_landscape:U_local 4 0.5326 0.03556 0.8180 0.779
## Residual 48 7.8127 0.52169
## Total 80 14.9757 1.00000
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1knitr::kable(perman)| Df | SumOfSqs | R2 | F | Pr(>F) | |
|---|---|---|---|---|---|
| plotid | 26 | 5.9884451 | 0.3998777 | 1.4150825 | 0.142 |
| U_local | 2 | 0.6419672 | 0.0428673 | 1.9720772 | 0.018 |
| U_landscape:U_local | 4 | 0.5325962 | 0.0355640 | 0.8180486 | 0.779 |
| Residual | 48 | 7.8126829 | 0.5216910 | NA | NA |
| Total | 80 | 14.9756915 | 1.0000000 | NA | NA |
bd_local <- betadisper(ST.dist, info$U_local)
bd_local ##
## Homogeneity of multivariate dispersions
##
## Call: betadisper(d = ST.dist, group = info$U_local)
##
## No. of Positive Eigenvalues: 43
## No. of Negative Eigenvalues: 37
##
## Average distance to median:
## LOW MEDIUM HIGH
## 0.4067 0.3960 0.4131
##
## Eigenvalues for PCoA axes:
## (Showing 8 of 80 eigenvalues)
## PCoA1 PCoA2 PCoA3 PCoA4 PCoA5 PCoA6 PCoA7 PCoA8
## 4.3459 1.4790 1.3238 1.1250 0.8922 0.7949 0.6316 0.5680print(permutest(bd_local,pairwise=TRUE,permutations=999)) ## als het niet significant is, dan is het goed##
## Permutation test for homogeneity of multivariate dispersions
## Permutation: free
## Number of permutations: 999
##
## Response: Distances
## Df Sum Sq Mean Sq F N.Perm Pr(>F)
## Groups 2 0.00405 0.0020243 0.1404 999 0.868
## Residuals 78 1.12432 0.0144143
##
## Pairwise comparisons:
## (Observed p-value below diagonal, permuted p-value above diagonal)
## LOW MEDIUM HIGH
## LOW 0.76200 0.852
## MEDIUM 0.75211 0.595
## HIGH 0.84450 0.58989bd_land <- betadisper(ST.dist, info$U_landscape)
print(permutest(bd_land,pairwise=TRUE,permutations=999))##
## Permutation test for homogeneity of multivariate dispersions
## Permutation: free
## Number of permutations: 999
##
## Response: Distances
## Df Sum Sq Mean Sq F N.Perm Pr(>F)
## Groups 2 0.02294 0.011472 0.9652 999 0.385
## Residuals 78 0.92706 0.011885
##
## Pairwise comparisons:
## (Observed p-value below diagonal, permuted p-value above diagonal)
## LOW MEDIUM HIGH
## LOW 0.35600 0.117
## MEDIUM 0.37465 0.736
## HIGH 0.11768 0.72640no signs of “dispersion” => betadisper test is ok
info$interaction_comb <- interaction(info$U_landscape, info$U_local)
info$interaction_comb <- as.factor(info$interaction_comb)
pairwise.adonis(ST.dist, info$U_local) |> as_tibble() |> knitr::kable()| pairs | Df | SumsOfSqs | F.Model | R2 | p.value | p.adjusted | sig |
|---|---|---|---|---|---|---|---|
| LOW vs MEDIUM | 1 | 0.3239383 | 1.7890322 | 0.0332602 | 0.058 | 0.174 | |
| LOW vs HIGH | 1 | 0.5164513 | 2.7520290 | 0.0502635 | 0.007 | 0.021 | . |
| MEDIUM vs HIGH | 1 | 0.1225613 | 0.6713256 | 0.0127456 | 0.749 | 1.000 |
pairwise.adonis(ST.dist, info$U_landscape)## pairs Df SumsOfSqs F.Model R2 p.value p.adjusted sig
## 1 HIGH vs MEDIUM 1 0.1015874 0.5156143 0.009818305 0.898 1.000
## 2 HIGH vs LOW 1 0.3340548 1.8474924 0.034309721 0.053 0.159
## 3 MEDIUM vs LOW 1 0.2306406 1.2737888 0.023910235 0.200 0.600for the order hemiptera
atypical sample P24SG
data.st.dip <- data.st %>%
dplyr::filter(location != "P24SG" & order == "HEMIPTERA") %>%
droplevels()
dip.matrix <- data.st.dip %>%
dplyr::group_by(location, plotid, U_landscape, U_local, urb_cat, func_species) %>%
dplyr::summarise(total_n_indiv=sum(n_indiv)) %>%
spread(key=func_species, value=total_n_indiv, fill=0) %>%
subset(select=-c(plotid, U_landscape, U_local, urb_cat)) %>%
column_to_rownames(var="location")
##dit is de community matrix
##sum(dip.matrix) gives now 5287
info <- data.st.dip %>%
dplyr::group_by(location, plotid, U_landscape, U_local, urb_cat, func_species) %>%
dplyr::summarise(total_n_indiv=sum(n_indiv)) %>%
spread(key=func_species, value=total_n_indiv, fill=0) %>%
select(location, plotid, U_landscape, U_local, urb_cat)
## dit is de info
nmds_matrix=metaMDS(dip.matrix, distance="bray")## Square root transformation
## Wisconsin double standardization
## Run 0 stress 0.2880867
## Run 1 stress 0.2848808
## ... New best solution
## ... Procrustes: rmse 0.06113589 max resid 0.2215116
## Run 2 stress 0.2902554
## Run 3 stress 0.2967525
## Run 4 stress 0.2825922
## ... New best solution
## ... Procrustes: rmse 0.06269567 max resid 0.3127576
## Run 5 stress 0.2802908
## ... New best solution
## ... Procrustes: rmse 0.05111893 max resid 0.2331626
## Run 6 stress 0.2837882
## Run 7 stress 0.2818612
## Run 8 stress 0.2890104
## Run 9 stress 0.2854956
## Run 10 stress 0.2974851
## Run 11 stress 0.2834486
## Run 12 stress 0.2917573
## Run 13 stress 0.2889322
## Run 14 stress 0.2921012
## Run 15 stress 0.2930148
## Run 16 stress 0.2817663
## Run 17 stress 0.2832978
## Run 18 stress 0.2819635
## Run 19 stress 0.2955729
## Run 20 stress 0.2804275
## ... Procrustes: rmse 0.04593515 max resid 0.2261187
## *** Best solution was not repeated -- monoMDS stopping criteria:
## 4: no. of iterations >= maxit
## 16: stress ratio > sratmaxnmds_matrix##
## Call:
## metaMDS(comm = dip.matrix, distance = "bray")
##
## global Multidimensional Scaling using monoMDS
##
## Data: wisconsin(sqrt(dip.matrix))
## Distance: bray
##
## Dimensions: 2
## Stress: 0.2802908
## Stress type 1, weak ties
## Best solution was not repeated after 20 tries
## The best solution was from try 5 (random start)
## Scaling: centring, PC rotation, halfchange scaling
## Species: expanded scores based on 'wisconsin(sqrt(dip.matrix))'sppr<- specnumber(dip.matrix)
sppr## P01SG P01SY P01SR P02SG P02SY P02SR P03SG P03SY P03SR P04SG P04SY P04SR P05SG
## 6 5 11 4 9 4 5 7 7 4 9 6 13
## P05SY P05SR P06SG P06SY P06SR P07SG P07SY P07SR P08SG P08SY P08SR P09SG P09SY
## 11 7 8 6 4 9 9 7 8 7 4 9 5
## P09SR P10SG P10SY P10SR P11SG P11SY P11SR P12SG P12SY P12SR P13SG P13SY P13SR
## 4 14 7 9 12 5 5 8 11 8 11 3 10
## P14SG P14SY P14SR P15SG P15SY P15SR P16SG P16SY P16SR P17SG P17SY P17SR P18SG
## 10 7 6 5 12 5 8 9 12 5 5 8 10
## P18SY P18SR P19SG P19SY P19SR P20SG P20SY P20SR P21SG P21SY P21SR P22SG P22SY
## 7 5 9 7 6 5 8 6 9 5 15 7 5
## P22SR P23SG P23SY P23SR P24SY P24SR P25SG P25SY P25SR P26SG P26SY P26SR P27SG
## 13 8 6 5 2 5 5 8 3 9 6 5 4
## P27SY P27SR
## 4 3df_species_rich <- sppr %>%
enframe() %>%
full_join(info, by = c("name"="location")) %>%
rename("species_richness" = "value")
modA<-glmmTMB(species_richness ~ U_landscape+U_local+(1|plotid), data=df_species_rich, family=poisson) #poisson
summary(modA)## Family: poisson ( log )
## Formula: species_richness ~ U_landscape + U_local + (1 | plotid)
## Data: df_species_rich
##
## AIC BIC logLik deviance df.resid
## 392.1 406.4 -190.0 380.1 74
##
## Random effects:
##
## Conditional model:
## Groups Name Variance Std.Dev.
## plotid (Intercept) 4.72e-09 6.87e-05
## Number of obs: 80, groups: plotid, 27
##
## Conditional model:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.98105 0.09310 21.279 <2e-16 ***
## U_landscapeMEDIUM 0.09624 0.10461 0.920 0.358
## U_landscapeHIGH 0.15094 0.10222 1.477 0.140
## U_localMEDIUM -0.14086 0.10143 -1.389 0.165
## U_localHIGH -0.15173 0.10172 -1.492 0.136
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Anova(modA, type="3") #U_local is significant## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: species_richness
## Chisq Df Pr(>Chisq)
## (Intercept) 452.7930 1 <2e-16 ***
## U_landscape 2.2090 2 0.3314
## U_local 2.8254 2 0.2435
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot(simulateResiduals(modA))
testDispersion(simulateResiduals(modA))
##
## DHARMa nonparametric dispersion test via sd of residuals fitted vs.
## simulated
##
## data: simulationOutput
## dispersion = 1.0533, p-value = 0.656
## alternative hypothesis: two.sidedplot(allEffects(modA))
ggplot(df_species_rich, aes(x = U_landscape, y = species_richness, fill = U_local)) +
geom_boxplot(position=position_dodge(width=0.75)) +
geom_point(position=position_dodge(width=0.75), colour="gray3")+
scale_colour_manual(values = urb.col3) +
scale_fill_manual(values=urb.col3)+
theme_bw()+
#theme(legend.position = "none",
labs(x = "urbanisation level at local scale",
y = "Number of functional species per subplot",
title = "Species richness funct species HEMIPTERA")
### graph resembles previous because those orders make up the majority of the species
## new way:
NMDS.vect <- scores(nmds_matrix, display = "species") %>%
as.data.frame()
NMDS.scores <- scores(nmds_matrix, display = "sites") %>%
as.data.frame()%>%
rownames_to_column("location") %>%
full_join(info, by="location")
stressplot(nmds_matrix)
fig_nmds <- ggplot() +
geom_point(data = NMDS.scores, aes(x = NMDS1, y = NMDS2 , color = U_landscape)) +
scale_color_manual(values = urb.col3) +
geom_vline(xintercept = c(0), color = "grey70", linetype = 2) +
geom_hline(yintercept = c(0), color = "grey70", linetype = 2) +
clean_background +
labs(x = "NMDS1",
y = "NMDS2",
title = "NMDS func groups")
fig_nmds
plot_nmds <- ggplot(NMDS.scores, aes(x = NMDS1, y = NMDS2, color = U_landscape, shape = U_landscape)) +
geom_point(size = 3, alpha = 0.8) +
scale_color_manual(values = urb.col3) +
stat_ellipse(linetype = 2, size = 1) +
clean_background +
labs(title = "NMDS")## Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
## ℹ Please use `linewidth` instead.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.plot_nmds
plot_nmds <- ggplot(NMDS.scores, aes(x = NMDS1, y = NMDS2, color = U_local, shape = U_local)) +
geom_point(size = 3, alpha = 0.8) +
scale_color_manual(values = urb.col3) +
stat_ellipse(linetype = 2, size = 1) +
clean_background +
labs(title = "NMDS")
plot_nmds
plot_nmds <- ggplot(NMDS.scores, aes(x = NMDS1, y = NMDS2, color = U_local, shape =U_landscape)) +
geom_point(size = 3, alpha = 0.8) +
scale_color_manual(values = urb.col3) +
clean_background +
stat_ellipse(linetype =2 , size=1)+
labs(title = "NMDS")
plot_nmds
# envfit() takes the output of metaMDS() and the species matrix you created
fit <- envfit(nmds_matrix, dip.matrix, perm = 999)
#extract p-values for each species
fit_pvals <- fit$vectors$pvals %>%
as.data.frame() %>%
rownames_to_column("func_species") %>%
dplyr::rename("pvals" = ".")
# extract coordinates for species, only keep species with p-val = 0.001
fit_spp <- fit %>%
scores(., display = "vectors") %>%
as.data.frame() %>%
rownames_to_column("func_species") %>%
full_join(., fit_pvals, by = "func_species") %>%
filter(pvals <= 0.001)
nmds_plot_new <- ggplot(NMDS.scores, aes(x = NMDS1, y = NMDS2)) +
coord_fixed() +
geom_point(aes(color = U_landscape, shape = U_landscape), size = 3, alpha = 0.8) +
stat_ellipse(aes(color = U_landscape)) +
scale_color_manual(values = urb.col3) +
geom_segment(data = fit_spp, aes(x = 0, xend = NMDS1, y = 0, yend = NMDS2),
arrow = arrow(length = unit(0.25, "cm")),
col = "black") +
geom_text(data = fit_spp, aes(label = func_species), position = position_jitter(width = 0.3, height=0.1)) +
clean_background
nmds_plot_new
nmds_plot_new <- ggplot(NMDS.scores, aes(x = NMDS1, y = NMDS2)) +
coord_fixed() +
geom_point(aes(color = U_local, shape = U_landscape), size = 3, alpha = 0.8) +
scale_colour_manual(name= "Urbanisation \nat local scale", values= urb.col3)+
scale_shape_manual(name= "Urbanisation \nat local scale", values=c(16, 17, 15))+
geom_segment(data = fit_spp, aes(x = 0, xend = NMDS1, y = 0, yend = NMDS2),
arrow = arrow(length = unit(0.25, "cm")),
col = "black") +
geom_text(data = fit_spp, aes(label = func_species), position = position_jitter(width = 0.2, height=0.1)) +
clean_background
nmds_plot_new
for the order hymenoptera
atypical sample P20SG
data.st.dip <- data.st %>%
dplyr::filter(location !="P20SG" & order == "HYMENOPTERA") %>%
droplevels()
dip.matrix <- data.st.dip %>%
dplyr::group_by(location, plotid, U_landscape, U_local, urb_cat, func_species) %>%
dplyr::summarise(total_n_indiv=sum(n_indiv)) %>%
spread(key=func_species, value=total_n_indiv, fill=0) %>%
subset(select=-c(plotid, U_landscape, U_local, urb_cat)) %>%
column_to_rownames(var="location")
##dit is de community matrix
##sum(dip.matrix) gives now 5287
info <- data.st.dip %>%
dplyr::group_by(location, plotid, U_landscape, U_local, urb_cat, func_species) %>%
dplyr::summarise(total_n_indiv=sum(n_indiv)) %>%
spread(key=func_species, value=total_n_indiv, fill=0) %>%
select(location, plotid, U_landscape, U_local, urb_cat)
## dit is de info
nmds_matrix=metaMDS(dip.matrix, distance="bray")## Square root transformation
## Wisconsin double standardization
## Run 0 stress 0.2456012
## Run 1 stress 0.259548
## Run 2 stress 0.2471578
## Run 3 stress 0.257863
## Run 4 stress 0.2569904
## Run 5 stress 0.2526869
## Run 6 stress 0.253721
## Run 7 stress 0.2455937
## ... New best solution
## ... Procrustes: rmse 0.003113757 max resid 0.02404952
## Run 8 stress 0.2659906
## Run 9 stress 0.2558501
## Run 10 stress 0.2577476
## Run 11 stress 0.2544416
## Run 12 stress 0.2506899
## Run 13 stress 0.246599
## Run 14 stress 0.2455937
## ... New best solution
## ... Procrustes: rmse 0.0001967132 max resid 0.0009612624
## ... Similar to previous best
## Run 15 stress 0.2566664
## Run 16 stress 0.2540408
## Run 17 stress 0.2524546
## Run 18 stress 0.2523057
## Run 19 stress 0.2503778
## Run 20 stress 0.2507785
## *** Best solution repeated 1 timesnmds_matrix##
## Call:
## metaMDS(comm = dip.matrix, distance = "bray")
##
## global Multidimensional Scaling using monoMDS
##
## Data: wisconsin(sqrt(dip.matrix))
## Distance: bray
##
## Dimensions: 2
## Stress: 0.2455937
## Stress type 1, weak ties
## Best solution was repeated 1 time in 20 tries
## The best solution was from try 14 (random start)
## Scaling: centring, PC rotation, halfchange scaling
## Species: expanded scores based on 'wisconsin(sqrt(dip.matrix))'sppr<- specnumber(dip.matrix)
sppr## P01SG P01SY P01SR P02SG P02SY P02SR P03SG P03SY P03SR P04SG P04SY P04SR P05SG
## 2 4 2 7 3 5 2 5 4 4 7 4 6
## P05SY P05SR P06SG P06SY P06SR P07SG P07SY P07SR P08SG P08SY P08SR P09SG P09SY
## 4 3 10 3 3 6 9 5 8 2 7 10 6
## P09SR P10SG P10SY P10SR P11SG P11SY P11SR P12SG P12SY P12SR P13SG P13SY P13SR
## 5 10 3 4 8 5 7 8 8 3 5 3 6
## P14SG P14SY P14SR P15SG P15SY P15SR P16SG P16SY P16SR P17SG P17SY P17SR P18SG
## 8 6 8 6 9 3 4 6 3 7 8 7 5
## P18SY P18SR P19SG P19SY P19SR P20SY P20SR P21SG P21SY P21SR P22SG P22SY P22SR
## 3 8 3 4 4 3 9 4 4 4 1 4 5
## P23SG P23SY P23SR P24SG P24SY P24SR P25SG P25SY P25SR P26SG P26SY P26SR P27SG
## 7 6 7 3 3 3 5 1 5 3 3 6 6
## P27SY P27SR
## 5 9df_species_rich <- sppr %>%
enframe() %>%
full_join(info, by = c("name"="location")) %>%
rename("species_richness" = "value")
modA<-glmmTMB(species_richness ~ U_landscape+U_local+(1|plotid), data=df_species_rich, family=nbinom2) #nbinom 2 to correct for overdispersion## Warning in finalizeTMB(TMBStruc, obj, fit, h, data.tmb.old): Model convergence
## problem; false convergence (8). See vignette('troubleshooting'),
## help('diagnose')summary(modA)## Family: nbinom2 ( log )
## Formula: species_richness ~ U_landscape + U_local + (1 | plotid)
## Data: df_species_rich
##
## AIC BIC logLik deviance df.resid
## 361.8 378.5 -173.9 347.8 73
##
## Random effects:
##
## Conditional model:
## Groups Name Variance Std.Dev.
## plotid (Intercept) 0.00168 0.04099
## Number of obs: 80, groups: plotid, 27
##
## Dispersion parameter for nbinom2 family (): 4.29e+07
##
## Conditional model:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.81972 0.10519 17.299 <2e-16 ***
## U_landscapeMEDIUM -0.10392 0.11938 -0.870 0.384
## U_landscapeHIGH -0.15392 0.12234 -1.258 0.208
## U_localMEDIUM -0.18834 0.12102 -1.556 0.120
## U_localHIGH -0.09803 0.11818 -0.830 0.407
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Anova(modA, type="3") #U_local is significant## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: species_richness
## Chisq Df Pr(>Chisq)
## (Intercept) 299.2621 1 <2e-16 ***
## U_landscape 1.6882 2 0.4299
## U_local 2.4331 2 0.2963
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot(simulateResiduals(modA))
testDispersion(simulateResiduals(modA))
##
## DHARMa nonparametric dispersion test via sd of residuals fitted vs.
## simulated
##
## data: simulationOutput
## dispersion = 0.92833, p-value = 0.696
## alternative hypothesis: two.sidedplot(allEffects(modA))## Warning in Effect.glmmTMB(predictors, mod, vcov. = vcov., ...): overriding
## variance function for effects/dev.resids: computed variances may be incorrect## Warning in Effect.glmmTMB(predictors, mod, vcov. = vcov., ...): overriding
## variance function for effects/dev.resids: computed variances may be incorrect
ggplot(df_species_rich, aes(x = U_landscape, y = species_richness, fill = U_local)) +
geom_boxplot(position=position_dodge(width=0.75)) +
geom_point(position=position_dodge(width=0.75), colour="gray3")+
scale_colour_manual(values = urb.col3) +
scale_fill_manual(values=urb.col3)+
theme_bw()+
#theme(legend.position = "none",
labs(x = "urbanisation level at local scale",
y = "Number of functional species per subplot",
title = "Species richness funct species HYMENOPTERA")
### graph resembles previous because those orders make up the majority of the species
## new way:
NMDS.vect <- scores(nmds_matrix, display = "species") %>%
as.data.frame()
NMDS.scores <- scores(nmds_matrix, display = "sites") %>%
as.data.frame()%>%
rownames_to_column("location") %>%
full_join(info, by="location")
stressplot(nmds_matrix)
fig_nmds <- ggplot() +
geom_point(data = NMDS.scores, aes(x = NMDS1, y = NMDS2 , color = U_landscape)) +
scale_color_manual(values = urb.col3) +
geom_vline(xintercept = c(0), color = "grey70", linetype = 2) +
geom_hline(yintercept = c(0), color = "grey70", linetype = 2) +
clean_background +
labs(x = "NMDS1",
y = "NMDS2",
title = "NMDS func groups")
fig_nmds
plot_nmds <- ggplot(NMDS.scores, aes(x = NMDS1, y = NMDS2, color = U_landscape, shape = U_landscape)) +
geom_point(size = 3, alpha = 0.8) +
scale_color_manual(values = urb.col3) +
stat_ellipse(linetype = 2, size = 1) +
clean_background +
labs(title = "NMDS")
plot_nmds
plot_nmds <- ggplot(NMDS.scores, aes(x = NMDS1, y = NMDS2, color = U_local, shape = U_local)) +
geom_point(size = 3, alpha = 0.8) +
scale_color_manual(values = urb.col3) +
stat_ellipse(linetype = 2, size = 1) +
clean_background +
labs(title = "NMDS")
plot_nmds
plot_nmds <- ggplot(NMDS.scores, aes(x = NMDS1, y = NMDS2, color = U_local, shape =U_landscape)) +
geom_point(size = 3, alpha = 0.8) +
scale_color_manual(values = urb.col3) +
clean_background +
stat_ellipse(linetype =2 , size=1)+
labs(title = "NMDS")
plot_nmds
# envfit() takes the output of metaMDS() and the species matrix you created
fit <- envfit(nmds_matrix, dip.matrix, perm = 999)
#extract p-values for each species
fit_pvals <- fit$vectors$pvals %>%
as.data.frame() %>%
rownames_to_column("func_species") %>%
dplyr::rename("pvals" = ".")
# extract coordinates for species, only keep species with p-val = 0.001
fit_spp <- fit %>%
scores(., display = "vectors") %>%
as.data.frame() %>%
rownames_to_column("func_species") %>%
full_join(., fit_pvals, by = "func_species") %>%
filter(pvals <= 0.001)
nmds_plot_new <- ggplot(NMDS.scores, aes(x = NMDS1, y = NMDS2)) +
coord_fixed() +
geom_point(aes(color = U_landscape, shape = U_landscape), size = 3, alpha = 0.8) +
stat_ellipse(aes(color = U_landscape)) +
scale_color_manual(values = urb.col3) +
geom_segment(data = fit_spp, aes(x = 0, xend = NMDS1, y = 0, yend = NMDS2),
arrow = arrow(length = unit(0.25, "cm")),
col = "black") +
geom_text(data = fit_spp, aes(label = func_species), position = position_jitter(width = 0.3, height=0.1)) +
clean_background
nmds_plot_new
nmds_plot_new <- ggplot(NMDS.scores, aes(x = NMDS1, y = NMDS2)) +
coord_fixed() +
geom_point(aes(color = U_local, shape = U_landscape), size = 3, alpha = 0.8) +
scale_colour_manual(name= "Urbanisation \nat local scale", values= urb.col3)+
scale_shape_manual(name= "Urbanisation \nat local scale", values=c(16, 17, 15))+
geom_segment(data = fit_spp, aes(x = 0, xend = NMDS1, y = 0, yend = NMDS2),
arrow = arrow(length = unit(0.25, "cm")),
col = "black") +
geom_text(data = fit_spp, aes(label = func_species), position = position_jitter(width = 0.2, height=0.1)) +
clean_background
nmds_plot_new
#biodiversity indices
data.st.corr <- data.st
#%>% filter(location!="P20SG" & location!="P24SG") %>%
# droplevels() ## dit verandert er niets aan dus heb ik niet uitgevoerd
func.matrix <- data.st.corr %>%
dplyr::group_by(location, plotid, U_landscape, U_local, urb_cat, func_species) %>%
dplyr::summarise(total_n_indiv=sum(n_indiv)) %>%
spread(key=func_species, value=total_n_indiv, fill=0) %>%
subset(select=-c(plotid, U_landscape, U_local, urb_cat)) %>%
column_to_rownames(var="location")
##dit is de community matrix
info <- data.st.corr %>%
dplyr::group_by(location, plotid, U_landscape, U_local, urb_cat, func_species) %>%
dplyr::summarise(total_n_indiv=sum(n_indiv)) %>%
spread(key=func_species, value=total_n_indiv, fill=0) %>%
select(location, plotid, U_landscape, U_local, urb_cat)
## dit is de info
sppr<-specnumber(func.matrix) ## will give number of prey order within each sample
shannondiv <- diversity(func.matrix, index=c("shannon"))
#Pielou's evenness
Pielou <- shannondiv/log(sppr)
df_div <- data.frame(sppr, shannondiv, Pielou)
library(tibble)
df_div <- rownames_to_column(df_div, var="location")
df_div <- left_join(df_div, info, by="location")
df_div$U_landscape <- factor(df_div$U_landscape, levels = c("LOW", "MEDIUM", "HIGH"))
df_div$U_local <- factor(df_div$U_local, levels = c("LOW", "MEDIUM", "HIGH"))
sp_rich<-ggplot(df_div, aes(U_local, sppr, fill= U_local)) +
geom_boxplot()+
geom_jitter(width=0.25)+
scale_fill_manual(values = urb.col3, name="urbanisation level \nat local scale") +
facet_grid(~U_landscape)+
ylab("Species Richness")+
xlab("")+
theme_bw()+
theme(axis.text = element_text(size=rel(1)), axis.title=element_text(size=rel(1.1)))
sp_rich
g1 <- ggplot_gtable(ggplot_build(sp_rich))
stript <- which(grepl('strip-t', g1$layout$name))
fills <- scales::alpha(urb.col3, 0.30)
k <- 1
for (i in stript) {
j <- which(grepl('rect', g1$grobs[[i]]$grobs[[1]]$childrenOrder))
g1$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g1)
sp_shannon<-ggplot(df_div, aes(U_local, shannondiv, fill= U_local)) +
geom_boxplot()+
geom_jitter(width=0.25)+
scale_fill_manual(values = urb.col3, name="urbanisation level \nat local scale") +
facet_grid(~U_landscape)+
ylab("shannon diversity")+
xlab("")+
theme_bw()+
theme(axis.text = element_text(size=rel(1)), axis.title=element_text(size=rel(1.1)))
sp_shannon
g_sh <- ggplot_gtable(ggplot_build(sp_shannon))
stript <- which(grepl('strip-t', g_sh$layout$name))
fills <- scales::alpha(urb.col3, 0.30)
k <- 1
for (i in stript) {
j <- which(grepl('rect', g_sh$grobs[[i]]$grobs[[1]]$childrenOrder))
g_sh$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g_sh)
sp_pi<-ggplot(df_div, aes(U_local, Pielou, fill= U_local)) +
geom_boxplot()+
geom_jitter(width=0.25)+
scale_fill_manual(values = urb.col3, name="urbanisation level \nat local scale") +
facet_grid(~U_landscape)+
ylab("Pielou Eveness")+
xlab("")+
theme_bw()+
theme(axis.text = element_text(size=rel(1)), axis.title=element_text(size=rel(1.1)))
sp_pi
g_pi <- ggplot_gtable(ggplot_build(sp_pi))
stript <- which(grepl('strip-t', g_pi$layout$name))
fills <- scales::alpha(urb.col3, 0.30)
k <- 1
for (i in stript) {
j <- which(grepl('rect', g_pi$grobs[[i]]$grobs[[1]]$childrenOrder))
g_pi$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
k <- k+1
}
ggarrange(g_pi)
fig_div<-ggarrange(g1, g_sh, g_pi, nrow=3)
fig_div
annotate_figure(fig_div,bottom=text_grob("facets = Urbanisation level at landscape scale", vjust=0.2, size=15))
d1<-ggplot(df_div, aes(U_landscape, sppr, fill= U_local)) +
geom_boxplot(position=position_dodge(width=0.75))+
geom_jitter(position=position_dodge(width=0.75), alpha=0.6, size=1.5, color="gray3")+
scale_fill_manual(values = urb.col3, name="urbanisation level \nat local scale") +
ylab("Species Richness")+
xlab("urbanisation level at landscape scale")+
theme_bw()+
theme(axis.text = element_text(size=rel(0.9)), axis.title=element_text(size=rel(1)))
d2<-ggplot(df_div, aes(U_landscape, shannondiv, fill= U_local)) +
geom_boxplot(position=position_dodge(width=0.75))+
geom_jitter(position=position_dodge(width=0.75), alpha=0.6, size=1.5, color="gray3")+
scale_fill_manual(values = urb.col3, name="urbanisation level \nat local scale") +
ylab("shannon diversity")+
xlab("urbanisation level at landscape scale")+
theme_bw()+
theme(axis.text = element_text(size=rel(0.9)), axis.title=element_text(size=rel(1)))
d3<-ggplot(df_div, aes(U_landscape, Pielou, fill= U_local)) +
geom_boxplot(position=position_dodge(width=0.75))+
geom_jitter(position=position_dodge(width=0.75), alpha=0.6, size=1.5, color="gray3")+
scale_fill_manual(values = urb.col3, name="urbanisation level \nat local scale") +
ylab("Pielou Eveness")+
xlab("urbanisation level at landscape scale")+
theme_bw()+
theme(axis.text = element_text(size=rel(0.9)), axis.title=element_text(size=rel(1)))
fig_d <- ggarrange(d1, d2, d3, nrow=3)
annotate_figure(fig_d,top=text_grob("Effect of urbanisation on functional diversity arthropod communities", size=15, hjust=0.65))
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